diff --git a/browser/ai_chat/BUILD.gn b/browser/ai_chat/BUILD.gn
index 80d9692b074..f9adf9cc9f3 100644
--- a/browser/ai_chat/BUILD.gn
+++ b/browser/ai_chat/BUILD.gn
@@ -66,6 +66,7 @@ static_library("ai_chat") {
     "//chrome/browser/actor",
     "//chrome/browser/profiles:profile",
     "//chrome/browser/ui/tabs:tabs_public",
+    "//chrome/common",
     "//chrome/common:channel_info",
     "//components/browsing_data/core",
     "//components/keyed_service/content",
@@ -76,6 +77,7 @@ static_library("ai_chat") {
     "//printing:printing_base",
     "//services/data_decoder/public/cpp",
     "//services/network/public/cpp",
+    "//third_party/blink/public/common",
     "//ui/base",
     "//ui/shell_dialogs",
   ]
@@ -84,7 +86,14 @@ static_library("ai_chat") {
     deps += [ "//brave/build/android:jni_headers" ]
   }
   if (!is_android) {
-    deps += [ "//chrome/browser/ui/side_panel" ]
+    sources += [
+      "text_file_extractor.cc",
+      "text_file_extractor.h",
+    ]
+    deps += [
+      "//chrome/browser/ui/side_panel",
+      "//third_party/blink/public/strings",
+    ]
   }
 
   if (enable_pdf) {
@@ -289,6 +298,10 @@ source_set("unit_tests") {
     sources += [ "tools/tab_management_tool_unittest.cc" ]
   }
 
+  if (!is_android) {
+    sources += [ "text_file_extractor_unittest.cc" ]
+  }
+
   if (enable_pdf) {
     sources += [ "pdf_text_extractor_unittest.cc" ]
   }
@@ -357,6 +370,7 @@ source_set("browser_tests") {
       "ai_chat_ui_browsertest.cc",
       "code_execution_tool_browsertest.cc",
       "page_content_fetcher_browsertest.cc",
+      "text_file_extractor_browsertest.cc",
     ]
 
     if (enable_pdf) {
diff --git a/browser/ai_chat/DEPS b/browser/ai_chat/DEPS
index d9a81c84e05..e3cea864cae 100644
--- a/browser/ai_chat/DEPS
+++ b/browser/ai_chat/DEPS
@@ -2,6 +2,7 @@ include_rules = [
   "+brave/components/restricted_web_contents_delegate",
   "+brave/services/printing/public/mojom",
   "+brave/components/text_recognition/common",
+  "+third_party/blink/public/strings",
 ]
 
 specific_include_rules = {
diff --git a/browser/ai_chat/file_text_extractor_base.cc b/browser/ai_chat/file_text_extractor_base.cc
index 2272b30ad58..f2b8bda0696 100644
--- a/browser/ai_chat/file_text_extractor_base.cc
+++ b/browser/ai_chat/file_text_extractor_base.cc
@@ -64,11 +64,34 @@ FileTextExtractorBase::~FileTextExtractorBase() {
   Cleanup();
 }
 
+void FileTextExtractorBase::ExtractText(
+    content::BrowserContext* browser_context,
+    const base::FilePath& file_path,
+    ExtractTextCallback callback) {
+  CHECK(!callback_) << "ExtractText called while extraction in progress";
+  callback_ = std::move(callback);
+  LoadInWebContents(browser_context, file_path);
+}
+
+void FileTextExtractorBase::ExtractText(
+    content::BrowserContext* browser_context,
+    std::vector<uint8_t> file_bytes,
+    const base::FilePath::StringType& extension,
+    ExtractTextCallback callback) {
+  CHECK(!callback_) << "ExtractText called while extraction in progress";
+  callback_ = std::move(callback);
+  WriteTempFileAndLoad(browser_context, std::move(file_bytes), extension);
+}
+
 network::mojom::WebSandboxFlags
 FileTextExtractorBase::AdditionalUnsandboxFlags() const {
   return network::mojom::WebSandboxFlags::kNone;
 }
 
+GURL FileTextExtractorBase::GetLoadURL(const base::FilePath& file_path) const {
+  return net::FilePathToFileURL(file_path);
+}
+
 void FileTextExtractorBase::LoadInWebContents(
     content::BrowserContext* browser_context,
     const base::FilePath& file_path) {
@@ -89,8 +112,8 @@ void FileTextExtractorBase::LoadInWebContents(
                        base::BindOnce(&FileTextExtractorBase::OnTimeout,
                                       base::Unretained(this)));
 
-  const GURL file_url = net::FilePathToFileURL(file_path);
-  web_contents_->GetController().LoadURL(file_url, content::Referrer(),
+  const GURL load_url = GetLoadURL(file_path);
+  web_contents_->GetController().LoadURL(load_url, content::Referrer(),
                                          ui::PAGE_TRANSITION_AUTO_TOPLEVEL,
                                          std::string());
 }
diff --git a/browser/ai_chat/file_text_extractor_base.h b/browser/ai_chat/file_text_extractor_base.h
index b800760cb39..209fead5220 100644
--- a/browser/ai_chat/file_text_extractor_base.h
+++ b/browser/ai_chat/file_text_extractor_base.h
@@ -18,6 +18,7 @@
 #include "brave/components/restricted_web_contents_delegate/restricted_web_contents_delegate.h"
 #include "content/public/browser/web_contents_observer.h"
 #include "services/network/public/cpp/web_sandbox_flags.h"
+#include "url/gurl.h"
 
 namespace content {
 class BrowserContext;
@@ -44,6 +45,21 @@ class FileTextExtractorBase : public RestrictedWebContentsDelegate,
   FileTextExtractorBase(const FileTextExtractorBase&) = delete;
   FileTextExtractorBase& operator=(const FileTextExtractorBase&) = delete;
 
+  // Two entry points for text extraction:
+
+  // Use an existing file path directly (e.g. from file picker).
+  void ExtractText(content::BrowserContext* browser_context,
+                   const base::FilePath& file_path,
+                   ExtractTextCallback callback);
+
+  // Write bytes to a temp file first (e.g. from drag-and-drop).
+  // |extension| is the file extension for MIME type detection (without
+  // leading dot).
+  void ExtractText(content::BrowserContext* browser_context,
+                   std::vector<uint8_t> file_bytes,
+                   const base::FilePath::StringType& extension,
+                   ExtractTextCallback callback);
+
  protected:
   // Called when the document has loaded and is ready for text extraction.
   // Subclasses must implement this and call Finish() with the result.
@@ -53,6 +69,10 @@ class FileTextExtractorBase : public RestrictedWebContentsDelegate,
   // (Scripts, Origin, Navigation). Override to add more, e.g. kPlugins.
   virtual network::mojom::WebSandboxFlags AdditionalUnsandboxFlags() const;
 
+  // Returns the URL to load for the given file path. Default returns a
+  // file:// URL. Override to customize, e.g. view-source:file://.
+  virtual GURL GetLoadURL(const base::FilePath& file_path) const;
+
   // Starts loading a file in a hidden WebContents.
   void LoadInWebContents(content::BrowserContext* browser_context,
                          const base::FilePath& file_path);
diff --git a/browser/ai_chat/pdf_text_extractor.cc b/browser/ai_chat/pdf_text_extractor.cc
index 9b0f8c85049..0647f2c3408 100644
--- a/browser/ai_chat/pdf_text_extractor.cc
+++ b/browser/ai_chat/pdf_text_extractor.cc
@@ -19,23 +19,6 @@ PdfTextExtractor::PdfTextExtractor() = default;
 
 PdfTextExtractor::~PdfTextExtractor() = default;
 
-void PdfTextExtractor::ExtractText(content::BrowserContext* browser_context,
-                                   const base::FilePath& pdf_path,
-                                   ExtractTextCallback callback) {
-  CHECK(!callback_) << "ExtractText called while extraction in progress";
-  callback_ = std::move(callback);
-  LoadInWebContents(browser_context, pdf_path);
-}
-
-void PdfTextExtractor::ExtractText(content::BrowserContext* browser_context,
-                                   std::vector<uint8_t> pdf_bytes,
-                                   ExtractTextCallback callback) {
-  CHECK(!callback_) << "ExtractText called while extraction in progress";
-  callback_ = std::move(callback);
-  WriteTempFileAndLoad(browser_context, std::move(pdf_bytes),
-                       FILE_PATH_LITERAL("pdf"));
-}
-
 network::mojom::WebSandboxFlags PdfTextExtractor::AdditionalUnsandboxFlags()
     const {
   // Plugins are required for the PDF viewer MimeHandlerView.
diff --git a/browser/ai_chat/pdf_text_extractor.h b/browser/ai_chat/pdf_text_extractor.h
index a7cc0fa84d4..22b61547b6f 100644
--- a/browser/ai_chat/pdf_text_extractor.h
+++ b/browser/ai_chat/pdf_text_extractor.h
@@ -6,16 +6,11 @@
 #ifndef BRAVE_BROWSER_AI_CHAT_PDF_TEXT_EXTRACTOR_H_
 #define BRAVE_BROWSER_AI_CHAT_PDF_TEXT_EXTRACTOR_H_
 
-#include <optional>
-#include <string>
-#include <vector>
-
 #include "base/memory/weak_ptr.h"
 #include "brave/browser/ai_chat/file_text_extractor_base.h"
 #include "services/network/public/cpp/web_sandbox_flags.h"
 
 namespace content {
-class BrowserContext;
 class RenderFrameHost;
 }  // namespace content
 
@@ -25,28 +20,12 @@ namespace ai_chat {
 // The PDF viewer extension + ScreenAI OCR pipeline runs, then page text is
 // extracted via PDFDocumentHelper::GetPageText().
 //
-// Two entry points:
-//   - ExtractText(browser_context, pdf_path, callback)
-//       Uses an existing file path directly (e.g. from file picker).
-//   - ExtractText(browser_context, pdf_bytes, callback)
-//       Writes bytes to a temp file first (e.g. from drag-and-drop).
-//
 // The extractor should be kept alive until the callback fires.
 class PdfTextExtractor : public FileTextExtractorBase {
  public:
   PdfTextExtractor();
   ~PdfTextExtractor() override;
 
-  // Use an existing file path directly (no temp file created).
-  void ExtractText(content::BrowserContext* browser_context,
-                   const base::FilePath& pdf_path,
-                   ExtractTextCallback callback);
-
-  // Write bytes to a temp file first, then extract.
-  void ExtractText(content::BrowserContext* browser_context,
-                   std::vector<uint8_t> pdf_bytes,
-                   ExtractTextCallback callback);
-
  private:
   // FileTextExtractorBase:
   void OnDocumentReady() override;
diff --git a/browser/ai_chat/pdf_text_extractor_browsertest.cc b/browser/ai_chat/pdf_text_extractor_browsertest.cc
index 8bd4cf0a24e..c26402dedb8 100644
--- a/browser/ai_chat/pdf_text_extractor_browsertest.cc
+++ b/browser/ai_chat/pdf_text_extractor_browsertest.cc
@@ -60,7 +60,7 @@ IN_PROC_BROWSER_TEST_F(PdfTextExtractorBrowserTest,
   base::test::TestFuture<std::optional<std::string>> future;
 
   extractor->ExtractText(browser_context(), std::move(*pdf_bytes),
-                         future.GetCallback());
+                         FILE_PATH_LITERAL("pdf"), future.GetCallback());
 
   auto result = future.Take();
   ASSERT_TRUE(result.has_value());
diff --git a/browser/ai_chat/pdf_text_extractor_unittest.cc b/browser/ai_chat/pdf_text_extractor_unittest.cc
index 16d6a470eca..31d13bf4515 100644
--- a/browser/ai_chat/pdf_text_extractor_unittest.cc
+++ b/browser/ai_chat/pdf_text_extractor_unittest.cc
@@ -40,7 +40,7 @@ TEST_F(PdfTextExtractorTest, BytesOverload_TimeoutReturnsNullopt) {
 
   std::vector<uint8_t> dummy_pdf = {0x25, 0x50, 0x44, 0x46};  // %PDF
   extractor->ExtractText(browser_context(), std::move(dummy_pdf),
-                         future.GetCallback());
+                         FILE_PATH_LITERAL("pdf"), future.GetCallback());
 
   // Fast-forward past the 30s extraction timeout.
   // This also processes pending ThreadPool tasks (temp-file write).
@@ -99,7 +99,7 @@ TEST_F(PdfTextExtractorTest, BytesOverload_CleanupAfterTimeout) {
 
   std::vector<uint8_t> dummy_pdf = {0x25, 0x50, 0x44, 0x46};
   extractor->ExtractText(browser_context(), std::move(dummy_pdf),
-                         future.GetCallback());
+                         FILE_PATH_LITERAL("pdf"), future.GetCallback());
 
   // Fast-forward past timeout to trigger cleanup.
   // This also processes pending ThreadPool tasks (temp-file write).
diff --git a/browser/ai_chat/text_file_extractor.cc b/browser/ai_chat/text_file_extractor.cc
new file mode 100644
index 00000000000..f71534bb52f
--- /dev/null
+++ b/browser/ai_chat/text_file_extractor.cc
@@ -0,0 +1,71 @@
+// Copyright (c) 2026 The Brave Authors. All rights reserved.
+// This Source Code Form is subject to the terms of the Mozilla Public
+// License, v. 2.0. If a copy of the MPL was not distributed with this file,
+// You can obtain one at https://mozilla.org/MPL/2.0/.
+
+#include "brave/browser/ai_chat/text_file_extractor.h"
+
+#include <utility>
+
+#include "base/functional/bind.h"
+#include "base/logging.h"
+#include "base/strings/strcat.h"
+#include "base/strings/string_util.h"
+#include "chrome/common/chrome_isolated_world_ids.h"
+#include "content/public/browser/render_frame_host.h"
+#include "content/public/browser/web_contents.h"
+#include "content/public/common/url_constants.h"
+#include "net/base/filename_util.h"
+#include "third_party/blink/public/strings/grit/blink_strings.h"
+#include "ui/base/l10n/l10n_util.h"
+
+namespace ai_chat {
+
+TextFileExtractor::TextFileExtractor() = default;
+
+TextFileExtractor::~TextFileExtractor() = default;
+
+GURL TextFileExtractor::GetLoadURL(const base::FilePath& file_path) const {
+  // Use view-source: to prevent HTML/XHTML from being rendered (which could
+  // execute scripts or load external resources). The source is displayed as
+  // raw text in the view-source document.
+  return GURL(base::StrCat({content::kViewSourceScheme, ":",
+                            net::FilePathToFileURL(file_path).spec()}));
+}
+
+void TextFileExtractor::OnDocumentReady() {
+  auto* rfh = GetWebContents()->GetPrimaryMainFrame();
+  if (!rfh) {
+    Finish(std::nullopt);
+    return;
+  }
+
+  rfh->ExecuteJavaScriptInIsolatedWorld(
+      u"document.body.innerText",
+      base::BindOnce(&TextFileExtractor::OnTextExtracted,
+                     weak_ptr_factory_.GetWeakPtr()),
+      ISOLATED_WORLD_ID_BRAVE_INTERNAL);
+}
+
+void TextFileExtractor::OnTextExtracted(base::Value result) {
+  if (result.is_string()) {
+    // view-source: innerText starts with the localized "Line wrap" label
+    // followed by a newline, and indents each line with a leading tab.
+    // Strip the label prefix and the per-line tabs.
+    // For non-empty files the prefix is "Line wrap\n\t", for empty files
+    // it's just "Line wrap" with no trailing newline or tab.
+    std::string text = std::move(result).TakeString();
+    std::string line_wrap_label =
+        l10n_util::GetStringUTF8(IDS_VIEW_SOURCE_LINE_WRAP);
+    if (auto rest = base::RemovePrefix(text, line_wrap_label)) {
+      text = std::string(base::RemovePrefix(*rest, "\n\t").value_or(*rest));
+    }
+    base::ReplaceSubstringsAfterOffset(&text, 0, "\n\t", "\n");
+    Finish(std::move(text));
+  } else {
+    DVLOG(1) << "TextFileExtractor: JS returned non-string result";
+    Finish(std::nullopt);
+  }
+}
+
+}  // namespace ai_chat
diff --git a/browser/ai_chat/text_file_extractor.h b/browser/ai_chat/text_file_extractor.h
new file mode 100644
index 00000000000..d573bf39d2c
--- /dev/null
+++ b/browser/ai_chat/text_file_extractor.h
@@ -0,0 +1,45 @@
+// Copyright (c) 2026 The Brave Authors. All rights reserved.
+// This Source Code Form is subject to the terms of the Mozilla Public
+// License, v. 2.0. If a copy of the MPL was not distributed with this file,
+// You can obtain one at https://mozilla.org/MPL/2.0/.
+
+#ifndef BRAVE_BROWSER_AI_CHAT_TEXT_FILE_EXTRACTOR_H_
+#define BRAVE_BROWSER_AI_CHAT_TEXT_FILE_EXTRACTOR_H_
+
+#include "base/gtest_prod_util.h"
+#include "base/memory/weak_ptr.h"
+#include "brave/browser/ai_chat/file_text_extractor_base.h"
+
+namespace ai_chat {
+
+// Extracts text from a file by loading it in a hidden background WebContents.
+// Chromium's renderer handles MIME sniffing and renders the file content,
+// then text is extracted via document.body.innerText.
+//
+// The extractor should be kept alive until the callback fires.
+class TextFileExtractor : public FileTextExtractorBase {
+ public:
+  TextFileExtractor();
+  ~TextFileExtractor() override;
+
+ private:
+  FRIEND_TEST_ALL_PREFIXES(TextFileExtractorTest,
+                           OnTextExtracted_StripsViewSourcePrefix);
+  FRIEND_TEST_ALL_PREFIXES(TextFileExtractorTest,
+                           OnTextExtracted_StripsLeadingTabs);
+  FRIEND_TEST_ALL_PREFIXES(TextFileExtractorTest, OnTextExtracted_EmptyFile);
+  FRIEND_TEST_ALL_PREFIXES(TextFileExtractorTest,
+                           OnTextExtracted_NonStringReturnsNullopt);
+
+  // FileTextExtractorBase:
+  GURL GetLoadURL(const base::FilePath& file_path) const override;
+  void OnDocumentReady() override;
+
+  void OnTextExtracted(base::Value result);
+
+  base::WeakPtrFactory<TextFileExtractor> weak_ptr_factory_{this};
+};
+
+}  // namespace ai_chat
+
+#endif  // BRAVE_BROWSER_AI_CHAT_TEXT_FILE_EXTRACTOR_H_
diff --git a/browser/ai_chat/text_file_extractor_browsertest.cc b/browser/ai_chat/text_file_extractor_browsertest.cc
new file mode 100644
index 00000000000..d72e0e48dd4
--- /dev/null
+++ b/browser/ai_chat/text_file_extractor_browsertest.cc
@@ -0,0 +1,139 @@
+// Copyright (c) 2026 The Brave Authors. All rights reserved.
+// This Source Code Form is subject to the terms of the Mozilla Public
+// License, v. 2.0. If a copy of the MPL was not distributed with this file,
+// You can obtain one at https://mozilla.org/MPL/2.0/.
+
+#include "brave/browser/ai_chat/text_file_extractor.h"
+
+#include <optional>
+#include <string>
+#include <vector>
+
+#include "base/files/file_util.h"
+#include "base/path_service.h"
+#include "base/test/test_future.h"
+#include "base/threading/thread_restrictions.h"
+#include "brave/components/constants/brave_paths.h"
+#include "chrome/browser/profiles/profile.h"
+#include "chrome/browser/ui/browser.h"
+#include "chrome/test/base/in_process_browser_test.h"
+#include "content/public/test/browser_test.h"
+#include "testing/gtest/include/gtest/gtest.h"
+
+namespace ai_chat {
+
+// Expected content of test/data/leo/dummy.txt
+constexpr char kExpectedTextContent[] =
+    "Hello from a text file.\nThis is line two.";
+
+class TextFileExtractorBrowserTest : public InProcessBrowserTest {
+ protected:
+  base::FilePath GetTestFilePath(std::string_view filename) {
+    base::ScopedAllowBlockingForTesting allow_blocking;
+    return base::PathService::CheckedGet(brave::DIR_TEST_DATA)
+        .AppendASCII("leo")
+        .AppendASCII(filename);
+  }
+
+  content::BrowserContext* browser_context() { return browser()->profile(); }
+};
+
+IN_PROC_BROWSER_TEST_F(TextFileExtractorBrowserTest,
+                       PathOverload_ExtractsText) {
+  base::FilePath txt_path = GetTestFilePath("dummy.txt");
+
+  auto extractor = std::make_unique<TextFileExtractor>();
+  base::test::TestFuture<std::optional<std::string>> future;
+
+  extractor->ExtractText(browser_context(), txt_path, future.GetCallback());
+
+  auto result = future.Take();
+  ASSERT_TRUE(result.has_value());
+  EXPECT_EQ(*result, kExpectedTextContent);
+}
+
+IN_PROC_BROWSER_TEST_F(TextFileExtractorBrowserTest,
+                       BytesOverload_ExtractsText) {
+  std::optional<std::vector<uint8_t>> file_bytes;
+  {
+    base::ScopedAllowBlockingForTesting allow_blocking;
+    file_bytes = base::ReadFileToBytes(GetTestFilePath("dummy.txt"));
+  }
+  ASSERT_TRUE(file_bytes.has_value());
+
+  auto extractor = std::make_unique<TextFileExtractor>();
+  base::test::TestFuture<std::optional<std::string>> future;
+
+  extractor->ExtractText(browser_context(), std::move(*file_bytes),
+                         FILE_PATH_LITERAL("txt"), future.GetCallback());
+
+  auto result = future.Take();
+  ASSERT_TRUE(result.has_value());
+  EXPECT_EQ(*result, kExpectedTextContent);
+}
+
+// HTML files must not be rendered — the extracted text should be the raw
+// source containing HTML tags, not the rendered text content.
+IN_PROC_BROWSER_TEST_F(TextFileExtractorBrowserTest,
+                       PathOverload_HtmlNotRendered) {
+  base::FilePath html_path = GetTestFilePath("dummy.html");
+
+  auto extractor = std::make_unique<TextFileExtractor>();
+  base::test::TestFuture<std::optional<std::string>> future;
+
+  extractor->ExtractText(browser_context(), html_path, future.GetCallback());
+
+  auto result = future.Take();
+  ASSERT_TRUE(result.has_value());
+  // Raw source must contain HTML tags — proves it was NOT rendered.
+  EXPECT_NE(result->find("<p>Hello from an HTML file.</p>"), std::string::npos);
+  EXPECT_NE(result->find("<script>"), std::string::npos);
+}
+
+IN_PROC_BROWSER_TEST_F(TextFileExtractorBrowserTest,
+                       BytesOverload_HtmlNotRendered) {
+  std::optional<std::vector<uint8_t>> file_bytes;
+  {
+    base::ScopedAllowBlockingForTesting allow_blocking;
+    file_bytes = base::ReadFileToBytes(GetTestFilePath("dummy.html"));
+  }
+  ASSERT_TRUE(file_bytes.has_value());
+
+  auto extractor = std::make_unique<TextFileExtractor>();
+  base::test::TestFuture<std::optional<std::string>> future;
+
+  extractor->ExtractText(browser_context(), std::move(*file_bytes),
+                         FILE_PATH_LITERAL("html"), future.GetCallback());
+
+  auto result = future.Take();
+  ASSERT_TRUE(result.has_value());
+  EXPECT_NE(result->find("<p>Hello from an HTML file.</p>"), std::string::npos);
+  EXPECT_NE(result->find("<script>"), std::string::npos);
+}
+
+IN_PROC_BROWSER_TEST_F(TextFileExtractorBrowserTest, EmptyFile_ReturnsEmpty) {
+  // Create a temporary empty file
+  base::FilePath temp_path;
+  {
+    base::ScopedAllowBlockingForTesting allow_blocking;
+    ASSERT_TRUE(base::CreateTemporaryFile(&temp_path));
+  }
+
+  auto extractor = std::make_unique<TextFileExtractor>();
+  base::test::TestFuture<std::optional<std::string>> future;
+
+  extractor->ExtractText(browser_context(), temp_path, future.GetCallback());
+
+  auto result = future.Take();
+  // Empty file should either return nullopt or empty string.
+  if (result.has_value()) {
+    EXPECT_TRUE(result->empty());
+  }
+
+  {
+    base::ScopedAllowBlockingForTesting allow_blocking;
+    base::DeleteFile(temp_path);
+  }
+}
+
+}  // namespace ai_chat
diff --git a/browser/ai_chat/text_file_extractor_unittest.cc b/browser/ai_chat/text_file_extractor_unittest.cc
new file mode 100644
index 00000000000..2e56e1c2d85
--- /dev/null
+++ b/browser/ai_chat/text_file_extractor_unittest.cc
@@ -0,0 +1,184 @@
+// Copyright (c) 2026 The Brave Authors. All rights reserved.
+// This Source Code Form is subject to the terms of the Mozilla Public
+// License, v. 2.0. If a copy of the MPL was not distributed with this file,
+// You can obtain one at https://mozilla.org/MPL/2.0/.
+
+#include "brave/browser/ai_chat/text_file_extractor.h"
+
+#include <optional>
+#include <string>
+#include <utility>
+#include <vector>
+
+#include "base/files/file_util.h"
+#include "base/files/scoped_temp_dir.h"
+#include "base/strings/strcat.h"
+#include "base/test/test_future.h"
+#include "chrome/test/base/testing_profile.h"
+#include "content/public/test/test_renderer_host.h"
+#include "testing/gtest/include/gtest/gtest.h"
+#include "third_party/blink/public/strings/grit/blink_strings.h"
+#include "ui/base/l10n/l10n_util.h"
+
+namespace ai_chat {
+
+class TextFileExtractorTest : public content::RenderViewHostTestHarness {
+ public:
+  TextFileExtractorTest()
+      : content::RenderViewHostTestHarness(
+            base::test::TaskEnvironment::TimeSource::MOCK_TIME) {}
+
+  std::unique_ptr<content::BrowserContext> CreateBrowserContext() override {
+    return std::make_unique<TestingProfile>();
+  }
+};
+
+// Without a real renderer producing text content, the extraction should
+// time out and return nullopt.
+TEST_F(TextFileExtractorTest, BytesOverload_TimeoutReturnsNullopt) {
+  auto extractor = std::make_unique<TextFileExtractor>();
+  base::test::TestFuture<std::optional<std::string>> future;
+
+  std::vector<uint8_t> text_bytes = {'h', 'e', 'l', 'l', 'o'};
+  extractor->ExtractText(browser_context(), std::move(text_bytes),
+                         FILE_PATH_LITERAL("txt"), future.GetCallback());
+
+  task_environment()->FastForwardBy(base::Seconds(31));
+
+  auto result = future.Take();
+  EXPECT_FALSE(result.has_value());
+}
+
+// Same timeout test but using the file-path overload (no temp file).
+TEST_F(TextFileExtractorTest, PathOverload_TimeoutReturnsNullopt) {
+  base::ScopedTempDir temp_dir;
+  ASSERT_TRUE(temp_dir.CreateUniqueTempDir());
+  base::FilePath txt_path = temp_dir.GetPath().AppendASCII("test.txt");
+  ASSERT_TRUE(base::WriteFile(txt_path, "hello world"));
+
+  auto extractor = std::make_unique<TextFileExtractor>();
+  base::test::TestFuture<std::optional<std::string>> future;
+
+  extractor->ExtractText(browser_context(), txt_path, future.GetCallback());
+
+  task_environment()->FastForwardBy(base::Seconds(31));
+
+  auto result = future.Take();
+  EXPECT_FALSE(result.has_value());
+}
+
+// Destroying the extractor while an extraction is in-flight must not crash
+// or leak.
+TEST_F(TextFileExtractorTest, DestroyDuringExtraction_NoCrash) {
+  auto extractor = std::make_unique<TextFileExtractor>();
+
+  base::ScopedTempDir temp_dir;
+  ASSERT_TRUE(temp_dir.CreateUniqueTempDir());
+  base::FilePath txt_path = temp_dir.GetPath().AppendASCII("test.txt");
+  ASSERT_TRUE(base::WriteFile(txt_path, "hello world"));
+
+  bool callback_called = false;
+  extractor->ExtractText(
+      browser_context(), txt_path,
+      base::BindOnce(
+          [](bool* called, std::optional<std::string>) { *called = true; },
+          &callback_called));
+
+  // Destroy while extraction is in progress — should not crash.
+  extractor.reset();
+
+  EXPECT_FALSE(callback_called);
+}
+
+// The bytes overload writes to a temp file. Verify cleanup completes
+// without crashing after timeout.
+TEST_F(TextFileExtractorTest, BytesOverload_CleanupAfterTimeout) {
+  auto extractor = std::make_unique<TextFileExtractor>();
+  base::test::TestFuture<std::optional<std::string>> future;
+
+  std::vector<uint8_t> text_bytes = {'h', 'e', 'l', 'l', 'o'};
+  extractor->ExtractText(browser_context(), std::move(text_bytes),
+                         FILE_PATH_LITERAL("txt"), future.GetCallback());
+
+  task_environment()->FastForwardBy(base::Seconds(31));
+
+  ASSERT_TRUE(future.Wait());
+}
+
+// The file-path overload should NOT delete the original file after extraction.
+TEST_F(TextFileExtractorTest, PathOverload_OriginalFileNotDeleted) {
+  base::ScopedTempDir temp_dir;
+  ASSERT_TRUE(temp_dir.CreateUniqueTempDir());
+  base::FilePath txt_path = temp_dir.GetPath().AppendASCII("keep_me.txt");
+  ASSERT_TRUE(base::WriteFile(txt_path, "this file should persist"));
+
+  auto extractor = std::make_unique<TextFileExtractor>();
+  base::test::TestFuture<std::optional<std::string>> future;
+
+  extractor->ExtractText(browser_context(), txt_path, future.GetCallback());
+
+  task_environment()->FastForwardBy(base::Seconds(31));
+
+  ASSERT_TRUE(future.Wait());
+
+  // The original file must still exist — only temp files are cleaned up.
+  EXPECT_TRUE(base::PathExists(txt_path));
+}
+
+// Unit tests for OnTextExtracted view-source stripping logic.
+TEST_F(TextFileExtractorTest, OnTextExtracted_StripsViewSourcePrefix) {
+  auto extractor = std::make_unique<TextFileExtractor>();
+  base::test::TestFuture<std::optional<std::string>> future;
+  extractor->callback_ = future.GetCallback();
+
+  // Simulate view-source innerText: "<Line wrap label>\n\tline1\n\tline2"
+  std::string line_wrap = l10n_util::GetStringUTF8(IDS_VIEW_SOURCE_LINE_WRAP);
+  extractor->OnTextExtracted(
+      base::Value(base::StrCat({line_wrap, "\n\tHello world\n\tSecond line"})));
+
+  auto result = future.Take();
+  ASSERT_TRUE(result.has_value());
+  EXPECT_EQ(*result, "Hello world\nSecond line");
+}
+
+TEST_F(TextFileExtractorTest, OnTextExtracted_StripsLeadingTabs) {
+  auto extractor = std::make_unique<TextFileExtractor>();
+  base::test::TestFuture<std::optional<std::string>> future;
+  extractor->callback_ = future.GetCallback();
+
+  // Tabs embedded within content (not at line start) should be preserved.
+  std::string line_wrap = l10n_util::GetStringUTF8(IDS_VIEW_SOURCE_LINE_WRAP);
+  extractor->OnTextExtracted(
+      base::Value(base::StrCat({line_wrap, "\n\tcol1\tcol2\n\tcol3\tcol4"})));
+
+  auto result = future.Take();
+  ASSERT_TRUE(result.has_value());
+  EXPECT_EQ(*result, "col1\tcol2\ncol3\tcol4");
+}
+
+TEST_F(TextFileExtractorTest, OnTextExtracted_EmptyFile) {
+  auto extractor = std::make_unique<TextFileExtractor>();
+  base::test::TestFuture<std::optional<std::string>> future;
+  extractor->callback_ = future.GetCallback();
+
+  // Empty file in view-source returns just the label with no newline.
+  std::string line_wrap = l10n_util::GetStringUTF8(IDS_VIEW_SOURCE_LINE_WRAP);
+  extractor->OnTextExtracted(base::Value(line_wrap));
+
+  auto result = future.Take();
+  ASSERT_TRUE(result.has_value());
+  EXPECT_TRUE(result->empty());
+}
+
+TEST_F(TextFileExtractorTest, OnTextExtracted_NonStringReturnsNullopt) {
+  auto extractor = std::make_unique<TextFileExtractor>();
+  base::test::TestFuture<std::optional<std::string>> future;
+  extractor->callback_ = future.GetCallback();
+
+  extractor->OnTextExtracted(base::Value(42));
+
+  auto result = future.Take();
+  EXPECT_FALSE(result.has_value());
+}
+
+}  // namespace ai_chat
diff --git a/browser/ai_chat/upload_file_helper.cc b/browser/ai_chat/upload_file_helper.cc
index b5e33387222..c112f8552c1 100644
--- a/browser/ai_chat/upload_file_helper.cc
+++ b/browser/ai_chat/upload_file_helper.cc
@@ -12,16 +12,18 @@
 #include "base/files/file.h"
 #include "base/files/file_path.h"
 #include "base/files/file_util.h"
-#include "base/no_destructor.h"
 #include "base/task/thread_pool.h"
 #include "brave/components/ai_chat/core/browser/utils.h"
 #include "brave/components/ai_chat/core/common/mojom/ai_chat.mojom.h"
 #include "brave/components/ai_chat/core/common/mojom/common.mojom.h"
 #include "chrome/browser/profiles/profile.h"
 #include "content/public/browser/web_contents.h"
+#include "media/base/mime_util.h"
+#include "net/base/mime_util.h"
 #include "printing/printing_utils.h"
 #include "services/data_decoder/public/cpp/data_decoder.h"
 #include "services/data_decoder/public/cpp/decode_image.h"
+#include "third_party/blink/public/common/mime_util/mime_util.h"
 #include "third_party/skia/include/core/SkBitmap.h"
 #include "ui/gfx/codec/png_codec.h"
 #include "ui/shell_dialogs/selected_file_info.h"
@@ -51,8 +53,29 @@ std::optional<mojom::UploadedFileType> DetermineFileType(
     return mojom::UploadedFileType::kImage;
   }
 
-  // If no recognized extension, return nullopt
-  return std::nullopt;
+  // If the extension has a known MIME type, reject clearly binary types
+  // (e.g. zip, exe, tar) while accepting text-renderable types.
+  auto extension = file_path.Extension();
+  if (!extension.empty()) {
+    std::string mime_type;
+    if (net::GetMimeTypeFromExtension(extension.substr(1), &mime_type)) {
+      // Reject known binary MIME types (media, archives, executables, etc.)
+      if (blink::IsSupportedImageMimeType(mime_type) ||
+          media::IsSupportedMediaMimeType(mime_type)) {
+        return std::nullopt;
+      }
+      // Accept types that Chromium can render as text
+      if (blink::IsSupportedNonImageMimeType(mime_type)) {
+        return mojom::UploadedFileType::kText;
+      }
+      // Known MIME type but not renderable (e.g. application/zip)
+      return std::nullopt;
+    }
+  }
+
+  // No MIME mapping for this extension (or no extension). Let the renderer
+  // try via MIME sniffing, matching how Chromium handles file:// URLs.
+  return mojom::UploadedFileType::kText;
 }
 
 // base::ReadFileToBytes doesn't handle content uri so we need to read from
@@ -129,9 +152,6 @@ void UploadFileHelper::UploadFile(std::unique_ptr<ui::SelectFilePolicy> policy,
   select_file_dialog_ = ui::SelectFileDialog::Create(this, std::move(policy));
   ui::SelectFileDialog::FileTypeInfo info;
   info.allowed_paths = ui::SelectFileDialog::FileTypeInfo::NATIVE_PATH;
-  info.extensions = {{FILE_PATH_LITERAL("png"), FILE_PATH_LITERAL("jpeg"),
-                      FILE_PATH_LITERAL("jpg"), FILE_PATH_LITERAL("webp"),
-                      FILE_PATH_LITERAL("pdf")}};
 #if BUILDFLAG(IS_ANDROID)
   // Set the list of acceptable MIME types for the file picker; this will apply
   // to any subsequent SelectFile() calls.
@@ -208,12 +228,16 @@ void UploadFileHelper::MultiFilesSelected(
                 uploaded_files.push_back(mojom::UploadedFile::New(
                     filepath.AsUTF8Unsafe(), file_data->size(), *file_data,
                     *file_type_opt, std::nullopt));
+              } else {
+                // Include rejected files as empty stubs so the frontend
+                // can detect dropped files and show an error alert.
+                uploaded_files.push_back(mojom::UploadedFile::New(
+                    filepath.AsUTF8Unsafe(), 0, std::vector<uint8_t>(),
+                    mojom::UploadedFileType::kText, std::nullopt));
               }
             }
             std::move(callback).Run(
-                uploaded_files.empty()
-                    ? std::nullopt
-                    : std::make_optional(std::move(uploaded_files)));
+                std::make_optional(std::move(uploaded_files)));
           },
           std::move(upload_file_callback_)));
 
@@ -241,8 +265,10 @@ void UploadFileHelper::MultiFilesSelected(
           auto file_type_opt = DetermineFileType(filepath, *file_data);
 
           if (file_type_opt &&
-              *file_type_opt == mojom::UploadedFileType::kPdf) {
-            // For PDFs, just return the raw data without processing
+              (*file_type_opt == mojom::UploadedFileType::kPdf ||
+               *file_type_opt == mojom::UploadedFileType::kText)) {
+            // For PDFs and text files, return raw data without processing.
+            // Text extraction happens via ProcessPdfFile/ProcessTextFile.
             std::move(callback).Run(std::make_tuple(
                 std::move(file_data), std::move(filepath), file_type_opt));
           } else if (file_type_opt &&
@@ -297,13 +323,14 @@ void UploadFileHelper::OnFileRead(
   // Determine file type based on extension and validate PDF content
   auto file_type_opt = DetermineFileType(std::get<1>(result), *file_data);
 
-  if (file_type_opt && *file_type_opt == mojom::UploadedFileType::kPdf) {
-    // Return raw PDF data; text extraction happens via ProcessPdfFile mojo
-    // endpoint when uploading from WebUI drag-and-drop.
+  if (file_type_opt && (*file_type_opt == mojom::UploadedFileType::kPdf ||
+                        *file_type_opt == mojom::UploadedFileType::kText)) {
+    // Return raw data; text extraction happens via ProcessPdfFile or
+    // ProcessTextFile mojo endpoint.
     std::vector<mojom::UploadedFilePtr> files;
-    files.push_back(mojom::UploadedFile::New(
-        std::get<1>(result).AsUTF8Unsafe(), file_data->size(), *file_data,
-        mojom::UploadedFileType::kPdf, std::nullopt));
+    files.push_back(mojom::UploadedFile::New(std::get<1>(result).AsUTF8Unsafe(),
+                                             file_data->size(), *file_data,
+                                             *file_type_opt, std::nullopt));
     std::move(upload_file_callback_).Run(std::make_optional(std::move(files)));
   } else if (file_type_opt &&
              *file_type_opt == mojom::UploadedFileType::kImage) {
@@ -314,8 +341,14 @@ void UploadFileHelper::OnFileRead(
                        weak_ptr_factory_.GetWeakPtr(),
                        std::get<1>(result).AsUTF8Unsafe()));
   } else {
-    // Fail if we cannot handle this file type
-    std::move(upload_file_callback_).Run(std::nullopt);
+    // Include as empty stub so the frontend can detect the unsupported
+    // file and show an error (as opposed to nullopt which means the user
+    // cancelled the file picker).
+    std::vector<mojom::UploadedFilePtr> files;
+    files.push_back(mojom::UploadedFile::New(
+        std::get<1>(result).AsUTF8Unsafe(), 0, std::vector<uint8_t>(),
+        mojom::UploadedFileType::kText, std::nullopt));
+    std::move(upload_file_callback_).Run(std::make_optional(std::move(files)));
   }
 }
 
diff --git a/browser/ai_chat/upload_file_helper_unittest.cc b/browser/ai_chat/upload_file_helper_unittest.cc
index 74add0409d5..83bf25a7c4b 100644
--- a/browser/ai_chat/upload_file_helper_unittest.cc
+++ b/browser/ai_chat/upload_file_helper_unittest.cc
@@ -135,18 +135,10 @@ TEST_F(UploadFileHelperTest, AcceptedFileExtensions) {
   EXPECT_CALL(observer, OnFilesSelected).Times(0);
   EXPECT_FALSE(UploadFileSync());
   EXPECT_EQ(dialog_params_.type, ui::SelectFileDialog::SELECT_OPEN_MULTI_FILE);
+  // No extension filtering — all files are accepted, matching how Chromium's
+  // Browser::OpenFile works. The renderer handles MIME sniffing.
   ASSERT_TRUE(dialog_params_.file_types);
-  ASSERT_EQ(1u, dialog_params_.file_types->extensions.size());
-  EXPECT_TRUE(std::ranges::contains(dialog_params_.file_types->extensions[0],
-                                    FILE_PATH_LITERAL("png")));
-  EXPECT_TRUE(std::ranges::contains(dialog_params_.file_types->extensions[0],
-                                    FILE_PATH_LITERAL("jpeg")));
-  EXPECT_TRUE(std::ranges::contains(dialog_params_.file_types->extensions[0],
-                                    FILE_PATH_LITERAL("jpg")));
-  EXPECT_TRUE(std::ranges::contains(dialog_params_.file_types->extensions[0],
-                                    FILE_PATH_LITERAL("webp")));
-  EXPECT_TRUE(std::ranges::contains(dialog_params_.file_types->extensions[0],
-                                    FILE_PATH_LITERAL("pdf")));
+  EXPECT_TRUE(dialog_params_.file_types->extensions.empty());
 #if BUILDFLAG(IS_ANDROID)
   // Android doesn't support view-source for text file extraction, so only
   // image and PDF MIME types are accepted.
@@ -302,8 +294,13 @@ TEST_F(UploadFileHelperTest, PdfFileWithInvalidHeader) {
   auto result = UploadFileSync();
   testing::Mock::VerifyAndClearExpectations(&observer);
 
-  // Should fail since it has .pdf extension but doesn't look like a PDF
-  EXPECT_FALSE(result);
+  // Should return a stub entry (empty data, kText) so the frontend can
+  // detect the rejection and show an error.
+  ASSERT_TRUE(result);
+  ASSERT_EQ(1u, result->size());
+  EXPECT_TRUE((*result)[0]->data.empty());
+  EXPECT_EQ((*result)[0]->type, mojom::UploadedFileType::kText);
+  EXPECT_FALSE((*result)[0]->extracted_text.has_value());
 }
 
 TEST_F(UploadFileHelperTest, PdfFileTooSmall) {
@@ -324,8 +321,117 @@ TEST_F(UploadFileHelperTest, PdfFileTooSmall) {
   auto result = UploadFileSync();
   testing::Mock::VerifyAndClearExpectations(&observer);
 
-  // Should fail since it has .pdf extension but is too small to be a valid PDF
-  EXPECT_FALSE(result);
+  // Should return a stub entry for the rejected file.
+  ASSERT_TRUE(result);
+  ASSERT_EQ(1u, result->size());
+  EXPECT_TRUE((*result)[0]->data.empty());
+  EXPECT_EQ((*result)[0]->type, mojom::UploadedFileType::kText);
+  EXPECT_FALSE((*result)[0]->extracted_text.has_value());
+}
+
+TEST_F(UploadFileHelperTest, BinaryFileRejected) {
+  // Test that files with known binary MIME types are rejected
+  base::FilePath zip_path = temp_dir_.GetPath().AppendASCII("archive.zip");
+  ASSERT_TRUE(base::WriteFile(zip_path, "PK\x03\x04 fake zip content"));
+
+  ui::SelectFileDialog::SetFactory(
+      std::make_unique<content::FakeSelectFileDialogFactory>(
+          std::vector<base::FilePath>{zip_path}));
+
+  testing::NiceMock<MockObserver> observer(file_helper_.get());
+  EXPECT_CALL(observer, OnFilesSelected).Times(1);
+
+  auto result = UploadFileSync();
+  testing::Mock::VerifyAndClearExpectations(&observer);
+
+  // Should return a stub entry for the rejected zip file.
+  ASSERT_TRUE(result);
+  ASSERT_EQ(1u, result->size());
+  EXPECT_TRUE((*result)[0]->data.empty());
+  EXPECT_EQ((*result)[0]->type, mojom::UploadedFileType::kText);
+  EXPECT_FALSE((*result)[0]->extracted_text.has_value());
+}
+
+TEST_F(UploadFileHelperTest, TextFileHandling) {
+  // Test text file with known extension
+  base::FilePath txt_path = temp_dir_.GetPath().AppendASCII("readme.txt");
+  ASSERT_TRUE(base::WriteFile(txt_path, "Hello, world!"));
+
+  ui::SelectFileDialog::SetFactory(
+      std::make_unique<content::FakeSelectFileDialogFactory>(
+          std::vector<base::FilePath>{txt_path}));
+
+  testing::NiceMock<MockObserver> observer(file_helper_.get());
+  EXPECT_CALL(observer, OnFilesSelected).Times(1);
+
+  auto result = UploadFileSync();
+  testing::Mock::VerifyAndClearExpectations(&observer);
+
+  ASSERT_TRUE(result);
+  ASSERT_EQ(1u, result->size());
+  EXPECT_EQ((*result)[0]->filename, ExpectedFilename(txt_path));
+  EXPECT_EQ((*result)[0]->type, mojom::UploadedFileType::kText);
+  EXPECT_EQ((*result)[0]->data.size(), 13u);
+}
+
+TEST_F(UploadFileHelperTest, TextFileWithoutExtension) {
+  // Test file without extension — should be treated as text
+  base::FilePath no_ext_path = temp_dir_.GetPath().AppendASCII("textfile");
+  ASSERT_TRUE(base::WriteFile(no_ext_path, "some text content"));
+
+  ui::SelectFileDialog::SetFactory(
+      std::make_unique<content::FakeSelectFileDialogFactory>(
+          std::vector<base::FilePath>{no_ext_path}));
+
+  testing::NiceMock<MockObserver> observer(file_helper_.get());
+  EXPECT_CALL(observer, OnFilesSelected).Times(1);
+
+  auto result = UploadFileSync();
+  testing::Mock::VerifyAndClearExpectations(&observer);
+
+  ASSERT_TRUE(result);
+  ASSERT_EQ(1u, result->size());
+  EXPECT_EQ((*result)[0]->type, mojom::UploadedFileType::kText);
+}
+
+TEST_F(UploadFileHelperTest, TextFileWithTrailingDot) {
+  // Test file ending with a dot — should be treated as text
+  base::FilePath dot_path = temp_dir_.GetPath().AppendASCII("file.");
+  ASSERT_TRUE(base::WriteFile(dot_path, "trailing dot content"));
+
+  ui::SelectFileDialog::SetFactory(
+      std::make_unique<content::FakeSelectFileDialogFactory>(
+          std::vector<base::FilePath>{dot_path}));
+
+  testing::NiceMock<MockObserver> observer(file_helper_.get());
+  EXPECT_CALL(observer, OnFilesSelected).Times(1);
+
+  auto result = UploadFileSync();
+  testing::Mock::VerifyAndClearExpectations(&observer);
+
+  ASSERT_TRUE(result);
+  ASSERT_EQ(1u, result->size());
+  EXPECT_EQ((*result)[0]->type, mojom::UploadedFileType::kText);
+}
+
+TEST_F(UploadFileHelperTest, TextFileWithUnknownExtension) {
+  // Test file with extension not in Chromium's MIME registry
+  base::FilePath diff_path = temp_dir_.GetPath().AppendASCII("changes.diff");
+  ASSERT_TRUE(base::WriteFile(diff_path, "--- a/file\n+++ b/file\n"));
+
+  ui::SelectFileDialog::SetFactory(
+      std::make_unique<content::FakeSelectFileDialogFactory>(
+          std::vector<base::FilePath>{diff_path}));
+
+  testing::NiceMock<MockObserver> observer(file_helper_.get());
+  EXPECT_CALL(observer, OnFilesSelected).Times(1);
+
+  auto result = UploadFileSync();
+  testing::Mock::VerifyAndClearExpectations(&observer);
+
+  ASSERT_TRUE(result);
+  ASSERT_EQ(1u, result->size());
+  EXPECT_EQ((*result)[0]->type, mojom::UploadedFileType::kText);
 }
 
 TEST_F(UploadFileHelperTest, MixedFileTypes) {
@@ -360,4 +466,42 @@ TEST_F(UploadFileHelperTest, MixedFileTypes) {
   EXPECT_GT((*result)[1]->data.size(), 0u);
 }
 
+TEST_F(UploadFileHelperTest, MixedWithUnsupportedFileDropsInvalid) {
+  data_decoder::test::InProcessDataDecoder data_decoder;
+  // Upload a valid image and text file alongside an unsupported zip file.
+  auto png_bytes = gfx::test::CreatePNGBytes(100);
+  base::FilePath png_path = temp_dir_.GetPath().AppendASCII("photo.png");
+  base::FilePath txt_path = temp_dir_.GetPath().AppendASCII("readme.txt");
+  base::FilePath zip_path = temp_dir_.GetPath().AppendASCII("archive.zip");
+  ASSERT_TRUE(base::WriteFile(png_path, base::span(*png_bytes)));
+  ASSERT_TRUE(base::WriteFile(txt_path, "hello world"));
+  ASSERT_TRUE(base::WriteFile(zip_path, "PK\x03\x04 fake zip"));
+
+  ui::SelectFileDialog::SetFactory(
+      std::make_unique<content::FakeSelectFileDialogFactory>(
+          std::vector<base::FilePath>{zip_path, png_path, txt_path}));
+
+  testing::NiceMock<MockObserver> observer(file_helper_.get());
+  EXPECT_CALL(observer, OnFilesSelected).Times(1);
+
+  auto result = UploadFileSync();
+  testing::Mock::VerifyAndClearExpectations(&observer);
+
+  // All 3 files returned: image and text are valid, zip is a stub.
+  ASSERT_TRUE(result);
+  ASSERT_EQ(3u, result->size());
+  // Barrier callback order is nondeterministic, so check by type.
+  EXPECT_TRUE(std::ranges::any_of(*result, [](const auto& f) {
+    return f->type == mojom::UploadedFileType::kImage && !f->data.empty();
+  }));
+  // The text file and zip stub both have kText type. Distinguish by data.
+  EXPECT_TRUE(std::ranges::any_of(*result, [](const auto& f) {
+    return f->type == mojom::UploadedFileType::kText && !f->data.empty();
+  }));
+  // Zip stub: kText with empty data
+  EXPECT_TRUE(std::ranges::any_of(*result, [](const auto& f) {
+    return f->type == mojom::UploadedFileType::kText && f->data.empty();
+  }));
+}
+
 }  // namespace ai_chat
diff --git a/browser/ui/webui/ai_chat/ai_chat_ui_page_handler.cc b/browser/ui/webui/ai_chat/ai_chat_ui_page_handler.cc
index 25e7534ae1c..302960ac0be 100644
--- a/browser/ui/webui/ai_chat/ai_chat_ui_page_handler.cc
+++ b/browser/ui/webui/ai_chat/ai_chat_ui_page_handler.cc
@@ -239,46 +239,70 @@ void AIChatUIPageHandler::UploadFile(bool use_media_capture,
 void AIChatUIPageHandler::OnFilesUploaded(
     UploadFileCallback callback,
     std::optional<std::vector<mojom::UploadedFilePtr>> uploaded_files) {
-#if BUILDFLAG(ENABLE_PDF)
   if (uploaded_files) {
-    // Collect PDF file paths and their indices before moving uploaded_files.
-    std::vector<std::pair<size_t, base::FilePath>> pdf_extractions;
+    // Collect all files needing text extraction in one pass.
+    // Store type alongside index/path so we can create the right extractor
+    // after uploaded_files is moved into the barrier callback.
+    struct ExtractionInfo {
+      size_t index;
+      base::FilePath path;
+      mojom::UploadedFileType type;
+    };
+    std::vector<ExtractionInfo> extractions;
     for (size_t i = 0; i < uploaded_files->size(); ++i) {
       auto& file = (*uploaded_files)[i];
-      if (file->type == mojom::UploadedFileType::kPdf &&
-          !file->extracted_text.has_value()) {
-        pdf_extractions.emplace_back(
-            i, base::FilePath::FromUTF8Unsafe(file->filename));
+      if (file->extracted_text.has_value() || file->data.empty()) {
+        continue;
+      }
+      bool needs_extraction = false;
+#if BUILDFLAG(ENABLE_PDF)
+      needs_extraction |= file->type == mojom::UploadedFileType::kPdf;
+#endif
+#if !BUILDFLAG(IS_ANDROID)
+      needs_extraction |= file->type == mojom::UploadedFileType::kText;
+#endif
+      if (needs_extraction) {
+        extractions.push_back(
+            {i, base::FilePath::FromUTF8Unsafe(file->filename), file->type});
       }
     }
 
-    if (!pdf_extractions.empty()) {
-      // Extract all PDFs in parallel via BarrierCallback.
+    if (!extractions.empty()) {
       auto barrier =
           base::BarrierCallback<std::pair<size_t, std::optional<std::string>>>(
-              pdf_extractions.size(),
-              base::BindOnce(&AIChatUIPageHandler::OnAllPdfTextsExtracted,
+              extractions.size(),
+              base::BindOnce(&AIChatUIPageHandler::OnAllFilesExtracted,
                              weak_ptr_factory_.GetWeakPtr(),
                              std::move(callback), std::move(uploaded_files)));
 
-      for (const auto& [idx, pdf_path] : pdf_extractions) {
-        auto extractor = std::make_unique<PdfTextExtractor>();
+      for (const auto& info : extractions) {
+        std::unique_ptr<FileTextExtractorBase> extractor;
+#if BUILDFLAG(ENABLE_PDF)
+        if (info.type == mojom::UploadedFileType::kPdf) {
+          extractor = std::make_unique<PdfTextExtractor>();
+        }
+#endif
+#if !BUILDFLAG(IS_ANDROID)
+        if (info.type == mojom::UploadedFileType::kText) {
+          extractor = std::make_unique<TextFileExtractor>();
+        }
+#endif
+        CHECK(extractor);
         auto* extractor_ptr = extractor.get();
-        pdf_extractors_.push_back(std::move(extractor));
+        extractors_.push_back(std::move(extractor));
         extractor_ptr->ExtractText(
-            profile_, pdf_path,
-            base::BindOnce(&AIChatUIPageHandler::OnSinglePdfTextExtracted,
-                           weak_ptr_factory_.GetWeakPtr(), extractor_ptr, idx,
-                           barrier));
+            profile_, info.path,
+            base::BindOnce(&AIChatUIPageHandler::OnSingleFileExtracted,
+                           weak_ptr_factory_.GetWeakPtr(), extractor_ptr,
+                           info.index, barrier));
       }
       return;
     }
   }
-#endif  // BUILDFLAG(ENABLE_PDF)
   FinishUpload(std::move(callback), std::move(uploaded_files));
 }
 
-void AIChatUIPageHandler::OnAllPdfTextsExtracted(
+void AIChatUIPageHandler::OnAllFilesExtracted(
     UploadFileCallback callback,
     std::optional<std::vector<mojom::UploadedFilePtr>> uploaded_files,
     std::vector<std::pair<size_t, std::optional<std::string>>> results) {
@@ -292,6 +316,18 @@ void AIChatUIPageHandler::OnAllPdfTextsExtracted(
   FinishUpload(std::move(callback), std::move(uploaded_files));
 }
 
+void AIChatUIPageHandler::OnSingleFileExtracted(
+    FileTextExtractorBase* extractor_ptr,
+    size_t file_index,
+    base::OnceCallback<void(std::pair<size_t, std::optional<std::string>>)>
+        barrier_cb,
+    std::optional<std::string> extracted_text) {
+  std::erase_if(extractors_, [extractor_ptr](const auto& e) {
+    return e.get() == extractor_ptr;
+  });
+  std::move(barrier_cb).Run({file_index, std::move(extracted_text)});
+}
+
 void AIChatUIPageHandler::FinishUpload(
     UploadFileCallback callback,
     std::optional<std::vector<mojom::UploadedFilePtr>> uploaded_files) {
@@ -327,20 +363,28 @@ void AIChatUIPageHandler::ProcessImageFile(
           filename, std::move(callback)));
 }
 
+void AIChatUIPageHandler::ProcessTextFile(const std::vector<uint8_t>& file_data,
+                                          const std::string& filename,
+                                          ProcessTextFileCallback callback) {
+#if !BUILDFLAG(IS_ANDROID)
+  const auto extension = base::FilePath::FromUTF8Unsafe(filename).Extension();
+  ExtractAndProcessFile(
+      std::make_unique<TextFileExtractor>(), file_data,
+      extension.size() > 1 ? extension.substr(1) : FILE_PATH_LITERAL("txt"),
+      filename, mojom::UploadedFileType::kText, std::move(callback));
+#else
+  // Android does not support background text extraction via view-source:.
+  std::move(callback).Run(nullptr);
+#endif
+}
+
 void AIChatUIPageHandler::ProcessPdfFile(const std::vector<uint8_t>& file_data,
                                          const std::string& filename,
                                          ProcessPdfFileCallback callback) {
 #if BUILDFLAG(ENABLE_PDF)
-  auto extractor = std::make_unique<PdfTextExtractor>();
-  auto* extractor_ptr = extractor.get();
-  pdf_extractors_.push_back(std::move(extractor));
-
-  auto response_data = std::vector<uint8_t>(file_data);
-  extractor_ptr->ExtractText(
-      profile_, std::vector<uint8_t>(file_data),
-      base::BindOnce(&AIChatUIPageHandler::OnPdfTextExtracted,
-                     weak_ptr_factory_.GetWeakPtr(), extractor_ptr, filename,
-                     std::move(response_data), std::move(callback)));
+  ExtractAndProcessFile(std::make_unique<PdfTextExtractor>(), file_data,
+                        FILE_PATH_LITERAL("pdf"), filename,
+                        mojom::UploadedFileType::kPdf, std::move(callback));
 #else
   auto uploaded_file = ai_chat::mojom::UploadedFile::New(
       filename, file_data.size(), file_data,
@@ -349,35 +393,46 @@ void AIChatUIPageHandler::ProcessPdfFile(const std::vector<uint8_t>& file_data,
 #endif  // BUILDFLAG(ENABLE_PDF)
 }
 
-#if BUILDFLAG(ENABLE_PDF)
-void AIChatUIPageHandler::OnSinglePdfTextExtracted(
-    PdfTextExtractor* extractor_ptr,
-    size_t file_index,
-    base::OnceCallback<void(std::pair<size_t, std::optional<std::string>>)>
-        barrier_cb,
-    std::optional<std::string> extracted_text) {
-  std::erase_if(pdf_extractors_, [extractor_ptr](const auto& e) {
-    return e.get() == extractor_ptr;
-  });
-  std::move(barrier_cb).Run({file_index, std::move(extracted_text)});
+void AIChatUIPageHandler::ExtractAndProcessFile(
+    std::unique_ptr<FileTextExtractorBase> extractor,
+    const std::vector<uint8_t>& file_data,
+    const base::FilePath::StringType& extension,
+    const std::string& filename,
+    mojom::UploadedFileType file_type,
+    base::OnceCallback<void(mojom::UploadedFilePtr)> callback) {
+  auto* extractor_ptr = extractor.get();
+  extractors_.push_back(std::move(extractor));
+  auto response_data = std::vector<uint8_t>(file_data);
+  extractor_ptr->ExtractText(
+      profile_, std::vector<uint8_t>(file_data), extension,
+      base::BindOnce(&AIChatUIPageHandler::OnFileExtracted,
+                     weak_ptr_factory_.GetWeakPtr(), extractor_ptr, filename,
+                     std::move(response_data), file_type, std::move(callback)));
 }
 
-void AIChatUIPageHandler::OnPdfTextExtracted(
-    PdfTextExtractor* extractor_ptr,
+void AIChatUIPageHandler::OnFileExtracted(
+    FileTextExtractorBase* extractor_ptr,
     std::string filename,
     std::vector<uint8_t> file_data,
-    ProcessPdfFileCallback callback,
+    mojom::UploadedFileType file_type,
+    base::OnceCallback<void(mojom::UploadedFilePtr)> callback,
     std::optional<std::string> extracted_text) {
-  auto file_size = file_data.size();
-  auto uploaded_file = ai_chat::mojom::UploadedFile::New(
-      std::move(filename), file_size, std::move(file_data),
-      ai_chat::mojom::UploadedFileType::kPdf, std::move(extracted_text));
-  std::move(callback).Run(std::move(uploaded_file));
-  std::erase_if(pdf_extractors_, [extractor_ptr](const auto& e) {
+  std::erase_if(extractors_, [extractor_ptr](const auto& e) {
     return e.get() == extractor_ptr;
   });
+  // Text extraction failure means the file is not text-renderable (e.g.
+  // binary). Return null so the frontend does not attach it.
+  if (!extracted_text.has_value() &&
+      file_type == mojom::UploadedFileType::kText) {
+    std::move(callback).Run(nullptr);
+    return;
+  }
+  auto file_size = file_data.size();
+  auto uploaded_file = ai_chat::mojom::UploadedFile::New(
+      std::move(filename), file_size, std::move(file_data), file_type,
+      std::move(extracted_text));
+  std::move(callback).Run(std::move(uploaded_file));
 }
-#endif  // BUILDFLAG(ENABLE_PDF)
 
 void AIChatUIPageHandler::GetPluralString(const std::string& key,
                                           int32_t count,
diff --git a/browser/ui/webui/ai_chat/ai_chat_ui_page_handler.h b/browser/ui/webui/ai_chat/ai_chat_ui_page_handler.h
index 3a525b6769c..7d853397c74 100644
--- a/browser/ui/webui/ai_chat/ai_chat_ui_page_handler.h
+++ b/browser/ui/webui/ai_chat/ai_chat_ui_page_handler.h
@@ -14,9 +14,15 @@
 #include "base/memory/weak_ptr.h"
 #include "base/scoped_observation.h"
 #include "base/task/cancelable_task_tracker.h"
+#include "brave/browser/ai_chat/file_text_extractor_base.h"
 #include "brave/browser/ai_chat/upload_file_helper.h"
+#include "build/build_config.h"
 #include "pdf/buildflags.h"
 
+#if !BUILDFLAG(IS_ANDROID)
+#include "brave/browser/ai_chat/text_file_extractor.h"
+#endif
+
 #if BUILDFLAG(ENABLE_PDF)
 #include "brave/browser/ai_chat/pdf_text_extractor.h"
 #endif
@@ -96,6 +102,9 @@ class AIChatUIPageHandler : public mojom::AIChatUIHandler,
   void ProcessPdfFile(const std::vector<uint8_t>& file_data,
                       const std::string& filename,
                       ProcessPdfFileCallback callback) override;
+  void ProcessTextFile(const std::vector<uint8_t>& file_data,
+                       const std::string& filename,
+                       ProcessTextFileCallback callback) override;
   void GetPluralString(const std::string& key,
                        int32_t count,
                        GetPluralStringCallback callback) override;
@@ -124,12 +133,16 @@ class AIChatUIPageHandler : public mojom::AIChatUIHandler,
  private:
   FRIEND_TEST_ALL_PREFIXES(AIChatUIPageHandlerBrowserTest,
                            OnFilesUploaded_WithPdf);
+#if !BUILDFLAG(IS_ANDROID)
+  FRIEND_TEST_ALL_PREFIXES(AIChatUIPageHandlerBrowserTest,
+                           OnFilesUploaded_WithText);
+#endif
   FRIEND_TEST_ALL_PREFIXES(AIChatUIPageHandlerTest,
                            FinishUpload_StripsPathToBasename);
   FRIEND_TEST_ALL_PREFIXES(AIChatUIPageHandlerTest,
                            OnFilesUploaded_NonPdfGoesToFinish);
   FRIEND_TEST_ALL_PREFIXES(AIChatUIPageHandlerTest,
-                           OnAllPdfTextsExtracted_AppliesResults);
+                           OnAllFilesExtracted_AppliesResults);
 
   class ChatContextObserver : public content::WebContentsObserver {
    public:
@@ -165,26 +178,37 @@ class AIChatUIPageHandler : public mojom::AIChatUIHandler,
   void OnFilesUploaded(
       UploadFileCallback callback,
       std::optional<std::vector<mojom::UploadedFilePtr>> uploaded_files);
-  void OnAllPdfTextsExtracted(
+  // Collects barrier results and applies extracted text to uploaded files.
+  void OnAllFilesExtracted(
       UploadFileCallback callback,
       std::optional<std::vector<mojom::UploadedFilePtr>> uploaded_files,
       std::vector<std::pair<size_t, std::optional<std::string>>> results);
-  void FinishUpload(
-      UploadFileCallback callback,
-      std::optional<std::vector<mojom::UploadedFilePtr>> uploaded_files);
-#if BUILDFLAG(ENABLE_PDF)
-  void OnSinglePdfTextExtracted(
-      PdfTextExtractor* extractor_ptr,
+  // Per-file barrier callback — removes the finished extractor and forwards.
+  void OnSingleFileExtracted(
+      FileTextExtractorBase* extractor_ptr,
       size_t file_index,
       base::OnceCallback<void(std::pair<size_t, std::optional<std::string>>)>
           barrier_cb,
       std::optional<std::string> extracted_text);
-  void OnPdfTextExtracted(PdfTextExtractor* extractor_ptr,
-                          std::string filename,
-                          std::vector<uint8_t> file_data,
-                          ProcessPdfFileCallback callback,
-                          std::optional<std::string> extracted_text);
-#endif  // BUILDFLAG(ENABLE_PDF)
+  void FinishUpload(
+      UploadFileCallback callback,
+      std::optional<std::vector<mojom::UploadedFilePtr>> uploaded_files);
+
+  // Shared helper for ProcessTextFile / ProcessPdfFile.
+  void ExtractAndProcessFile(
+      std::unique_ptr<FileTextExtractorBase> extractor,
+      const std::vector<uint8_t>& file_data,
+      const base::FilePath::StringType& extension,
+      const std::string& filename,
+      mojom::UploadedFileType file_type,
+      base::OnceCallback<void(mojom::UploadedFilePtr)> callback);
+  void OnFileExtracted(
+      FileTextExtractorBase* extractor_ptr,
+      std::string filename,
+      std::vector<uint8_t> file_data,
+      mojom::UploadedFileType file_type,
+      base::OnceCallback<void(mojom::UploadedFilePtr)> callback,
+      std::optional<std::string> extracted_text);
 
   raw_ptr<AIChatTabHelper> active_chat_tab_helper_ = nullptr;
   raw_ptr<content::WebContents> owner_web_contents_ = nullptr;
@@ -203,10 +227,8 @@ class AIChatUIPageHandler : public mojom::AIChatUIHandler,
   // DataDecoder instance for processing image data
   data_decoder::DataDecoder data_decoder_;
 
-#if BUILDFLAG(ENABLE_PDF)
-  // Active PDF text extractors (owned until extraction completes)
-  std::vector<std::unique_ptr<PdfTextExtractor>> pdf_extractors_;
-#endif
+  // Active file extractors (owned until extraction completes)
+  std::vector<std::unique_ptr<FileTextExtractorBase>> extractors_;
 
   mojo::Receiver<ai_chat::mojom::AIChatUIHandler> receiver_;
   mojo::Remote<ai_chat::mojom::ChatUI> chat_ui_;
diff --git a/browser/ui/webui/ai_chat/ai_chat_ui_page_handler_browsertest.cc b/browser/ui/webui/ai_chat/ai_chat_ui_page_handler_browsertest.cc
index 2245a61e202..901c08fa196 100644
--- a/browser/ui/webui/ai_chat/ai_chat_ui_page_handler_browsertest.cc
+++ b/browser/ui/webui/ai_chat/ai_chat_ui_page_handler_browsertest.cc
@@ -24,6 +24,7 @@
 #include "brave/components/ai_chat/core/common/features.h"
 #include "brave/components/ai_chat/core/common/mojom/tab_tracker.mojom.h"
 #include "brave/components/constants/brave_paths.h"
+#include "build/build_config.h"
 #include "chrome/browser/profiles/profile.h"
 #include "chrome/test/base/chrome_test_utils.h"
 #include "chrome/test/base/in_process_browser_test.h"
@@ -47,6 +48,11 @@
 
 namespace ai_chat {
 
+#if !BUILDFLAG(IS_ANDROID)
+constexpr char kExpectedTextContent[] =
+    "Hello from a text file.\nThis is line two.";
+#endif
+
 #if BUILDFLAG(ENABLE_PDF)
 constexpr char kExpectedPdfText[] = "This is the way\nI have spoken";
 #endif  // BUILDFLAG(ENABLE_PDF)
@@ -339,6 +345,103 @@ IN_PROC_BROWSER_TEST_F(AIChatUIPageHandlerBrowserTest,
 }
 #endif  // BUILDFLAG(ENABLE_PDF)
 
+#if !BUILDFLAG(IS_ANDROID)
+IN_PROC_BROWSER_TEST_F(AIChatUIPageHandlerBrowserTest, ProcessTextFile) {
+  auto* page_handler = GetPageHandler(web_contents());
+  ASSERT_TRUE(page_handler);
+
+  base::FilePath txt_path;
+  std::vector<uint8_t> txt_bytes;
+  {
+    base::ScopedAllowBlockingForTesting allow_blocking;
+    txt_path = base::PathService::CheckedGet(brave::DIR_TEST_DATA)
+                   .AppendASCII("leo")
+                   .AppendASCII("dummy.txt");
+    auto bytes = base::ReadFileToBytes(txt_path);
+    ASSERT_TRUE(bytes.has_value());
+    txt_bytes = std::move(*bytes);
+  }
+
+  base::test::TestFuture<ai_chat::mojom::UploadedFilePtr> future;
+  page_handler->ProcessTextFile(txt_bytes, "dummy.txt", future.GetCallback());
+
+  auto result = future.Take();
+  ASSERT_TRUE(result);
+  EXPECT_EQ(result->filename, "dummy.txt");
+  EXPECT_EQ(result->type, ai_chat::mojom::UploadedFileType::kText);
+  EXPECT_GT(result->data.size(), 0u);
+  ASSERT_TRUE(result->extracted_text.has_value());
+  EXPECT_EQ(*result->extracted_text, kExpectedTextContent);
+}
+
+// Verify that HTML files are not rendered (no script execution, no external
+// resource loading). The extracted text must be the raw HTML source.
+IN_PROC_BROWSER_TEST_F(AIChatUIPageHandlerBrowserTest, ProcessHtmlFile) {
+  auto* page_handler = GetPageHandler(web_contents());
+  ASSERT_TRUE(page_handler);
+
+  base::FilePath html_path;
+  std::vector<uint8_t> html_bytes;
+  {
+    base::ScopedAllowBlockingForTesting allow_blocking;
+    html_path = base::PathService::CheckedGet(brave::DIR_TEST_DATA)
+                    .AppendASCII("leo")
+                    .AppendASCII("dummy.html");
+    auto bytes = base::ReadFileToBytes(html_path);
+    ASSERT_TRUE(bytes.has_value());
+    html_bytes = std::move(*bytes);
+  }
+
+  base::test::TestFuture<ai_chat::mojom::UploadedFilePtr> future;
+  page_handler->ProcessTextFile(html_bytes, "dummy.html", future.GetCallback());
+
+  auto result = future.Take();
+  ASSERT_TRUE(result);
+  EXPECT_EQ(result->filename, "dummy.html");
+  ASSERT_TRUE(result->extracted_text.has_value());
+  // Raw source must contain HTML tags — proves it was NOT rendered.
+  EXPECT_TRUE(result->extracted_text->find("<p>Hello from an HTML file.</p>") !=
+              std::string::npos);
+  EXPECT_TRUE(result->extracted_text->find("<script>") != std::string::npos);
+}
+
+IN_PROC_BROWSER_TEST_F(AIChatUIPageHandlerBrowserTest,
+                       OnFilesUploaded_WithText) {
+  auto* page_handler = GetPageHandler(web_contents());
+  ASSERT_TRUE(page_handler);
+
+  base::FilePath txt_path;
+  std::vector<uint8_t> txt_bytes;
+  {
+    base::ScopedAllowBlockingForTesting allow_blocking;
+    txt_path = base::PathService::CheckedGet(brave::DIR_TEST_DATA)
+                   .AppendASCII("leo")
+                   .AppendASCII("dummy.txt");
+    auto bytes = base::ReadFileToBytes(txt_path);
+    ASSERT_TRUE(bytes.has_value());
+    txt_bytes = std::move(*bytes);
+  }
+
+  std::vector<ai_chat::mojom::UploadedFilePtr> files;
+  files.push_back(ai_chat::mojom::UploadedFile::New(
+      txt_path.AsUTF8Unsafe(), txt_bytes.size(), std::move(txt_bytes),
+      ai_chat::mojom::UploadedFileType::kText, std::nullopt));
+
+  base::test::TestFuture<
+      std::optional<std::vector<ai_chat::mojom::UploadedFilePtr>>>
+      future;
+  page_handler->OnFilesUploaded(future.GetCallback(),
+                                std::make_optional(std::move(files)));
+
+  auto result = future.Take();
+  ASSERT_TRUE(result.has_value());
+  ASSERT_EQ(result->size(), 1u);
+  EXPECT_EQ((*result)[0]->filename, "dummy.txt");
+  ASSERT_TRUE((*result)[0]->extracted_text.has_value());
+  EXPECT_EQ(*(*result)[0]->extracted_text, kExpectedTextContent);
+}
+#endif  // !BUILDFLAG(IS_ANDROID)
+
 IN_PROC_BROWSER_TEST_F(AIChatUIPageHandlerBrowserTest,
                        AssociateURLDoesNotCrashShutdown) {
   auto* ai_chat_contents = web_contents();
diff --git a/browser/ui/webui/ai_chat/ai_chat_ui_page_handler_unittest.cc b/browser/ui/webui/ai_chat/ai_chat_ui_page_handler_unittest.cc
index eb486c5ae37..42a38a0e98e 100644
--- a/browser/ui/webui/ai_chat/ai_chat_ui_page_handler_unittest.cc
+++ b/browser/ui/webui/ai_chat/ai_chat_ui_page_handler_unittest.cc
@@ -166,7 +166,7 @@ TEST_F(AIChatUIPageHandlerTest, OnFilesUploaded_NonPdfGoesToFinish) {
   EXPECT_EQ((*result)[0]->filename, "photo.png");
 }
 
-TEST_F(AIChatUIPageHandlerTest, OnAllPdfTextsExtracted_AppliesResults) {
+TEST_F(AIChatUIPageHandlerTest, OnAllFilesExtracted_AppliesResults) {
   std::vector<mojom::UploadedFilePtr> files;
   files.push_back(
       mojom::UploadedFile::New("/path/doc1.pdf", 100, std::vector<uint8_t>(100),
@@ -177,30 +177,45 @@ TEST_F(AIChatUIPageHandlerTest, OnAllPdfTextsExtracted_AppliesResults) {
   files.push_back(
       mojom::UploadedFile::New("/path/doc2.pdf", 200, std::vector<uint8_t>(200),
                                mojom::UploadedFileType::kPdf, std::nullopt));
+  files.push_back(mojom::UploadedFile::New(
+      "/path/config.conf", 80, std::vector<uint8_t>(80),
+      mojom::UploadedFileType::kText, std::nullopt));
+  files.push_back(mojom::UploadedFile::New(
+      "/path/script.py", 120, std::vector<uint8_t>(120),
+      mojom::UploadedFileType::kText, std::nullopt));
 
   std::vector<std::pair<size_t, std::optional<std::string>>> results;
   results.emplace_back(0, "Text from doc1");
   results.emplace_back(2, std::nullopt);  // extraction failed for doc2
+  results.emplace_back(3, "key=value");
+  results.emplace_back(4, std::nullopt);  // extraction failed for script.py
 
   base::test::TestFuture<std::optional<std::vector<mojom::UploadedFilePtr>>>
       future;
-  page_handler()->OnAllPdfTextsExtracted(future.GetCallback(),
-                                         std::make_optional(std::move(files)),
-                                         std::move(results));
+  page_handler()->OnAllFilesExtracted(future.GetCallback(),
+                                      std::make_optional(std::move(files)),
+                                      std::move(results));
 
   auto result = future.Take();
   ASSERT_TRUE(result.has_value());
-  ASSERT_EQ(result->size(), 3u);
+  ASSERT_EQ(result->size(), 5u);
   // PDF extracted text applied and path stripped
   EXPECT_EQ((*result)[0]->filename, "doc1.pdf");
   ASSERT_TRUE((*result)[0]->extracted_text.has_value());
   EXPECT_EQ(*(*result)[0]->extracted_text, "Text from doc1");
-  // Non-PDF unaffected, path stripped
+  // Image unaffected, path stripped
   EXPECT_EQ((*result)[1]->filename, "photo.png");
   EXPECT_FALSE((*result)[1]->extracted_text.has_value());
-  // Failed extraction, path stripped
+  // Failed PDF extraction, path stripped
   EXPECT_EQ((*result)[2]->filename, "doc2.pdf");
   EXPECT_FALSE((*result)[2]->extracted_text.has_value());
+  // Text extracted and path stripped
+  EXPECT_EQ((*result)[3]->filename, "config.conf");
+  ASSERT_TRUE((*result)[3]->extracted_text.has_value());
+  EXPECT_EQ(*(*result)[3]->extracted_text, "key=value");
+  // Failed text extraction, path stripped
+  EXPECT_EQ((*result)[4]->filename, "script.py");
+  EXPECT_FALSE((*result)[4]->extracted_text.has_value());
 }
 
 }  // namespace ai_chat
diff --git a/components/ai_chat/core/browser/engine/oai_message_utils.cc b/components/ai_chat/core/browser/engine/oai_message_utils.cc
index 50eaf736037..acca7d0b5cc 100644
--- a/components/ai_chat/core/browser/engine/oai_message_utils.cc
+++ b/components/ai_chat/core/browser/engine/oai_message_utils.cc
@@ -310,6 +310,7 @@ std::vector<OAIMessage> BuildOAIMessages(
       std::vector<mojom::ContentBlockPtr> uploaded_images_content_blocks;
       std::vector<mojom::ContentBlockPtr> screenshots_content_blocks;
       std::vector<mojom::ContentBlockPtr> uploaded_pdfs_content_blocks;
+      std::vector<mojom::ContentBlockPtr> uploaded_text_files_content_blocks;
 
       uploaded_images_content_blocks.push_back(
           mojom::ContentBlock::NewTextContentBlock(mojom::TextContentBlock::New(
@@ -320,6 +321,9 @@ std::vector<OAIMessage> BuildOAIMessages(
       uploaded_pdfs_content_blocks.push_back(
           mojom::ContentBlock::NewTextContentBlock(mojom::TextContentBlock::New(
               "These PDFs are uploaded by the user")));
+      uploaded_text_files_content_blocks.push_back(
+          mojom::ContentBlock::NewTextContentBlock(mojom::TextContentBlock::New(
+              "These text files are uploaded by the user")));
 
       for (const auto& uploaded_file : *message->uploaded_files) {
         if (uploaded_file->type == mojom::UploadedFileType::kImage ||
@@ -354,6 +358,18 @@ std::vector<OAIMessage> BuildOAIMessages(
                             EngineConsumer::GetPdfDataURL(uploaded_file->data)),
                         pdf_filename)));
           }
+        } else if (uploaded_file->type == mojom::UploadedFileType::kText) {
+          std::string text_filename = uploaded_file->filename.empty()
+                                          ? "uploaded.txt"
+                                          : uploaded_file->filename;
+          if (uploaded_file->extracted_text.has_value() &&
+              !uploaded_file->extracted_text->empty()) {
+            uploaded_text_files_content_blocks.push_back(
+                mojom::ContentBlock::NewTextContentBlock(
+                    mojom::TextContentBlock::New(
+                        "[File: " + text_filename + "]\n" +
+                        *uploaded_file->extracted_text)));
+          }
         }
       }
 
@@ -377,6 +393,13 @@ std::vector<OAIMessage> BuildOAIMessages(
             std::make_move_iterator(uploaded_pdfs_content_blocks.begin()),
             std::make_move_iterator(uploaded_pdfs_content_blocks.end()));
       }
+
+      if (uploaded_text_files_content_blocks.size() > 1) {
+        oai_message.content.insert(
+            oai_message.content.end(),
+            std::make_move_iterator(uploaded_text_files_content_blocks.begin()),
+            std::make_move_iterator(uploaded_text_files_content_blocks.end()));
+      }
     }
 
     if (message->selected_text.has_value() &&
diff --git a/components/ai_chat/core/browser/engine/oai_message_utils_unittest.cc b/components/ai_chat/core/browser/engine/oai_message_utils_unittest.cc
index 23d477a2de7..8a965b5db25 100644
--- a/components/ai_chat/core/browser/engine/oai_message_utils_unittest.cc
+++ b/components/ai_chat/core/browser/engine/oai_message_utils_unittest.cc
@@ -585,6 +585,60 @@ TEST_F(OAIMessageUtilsTest, BuildOAIMessages_PdfExtractedTextPreferred) {
   VerifyTextBlock(FROM_HERE, messages[0].content[3], "query0");
 }
 
+TEST_F(OAIMessageUtilsTest, BuildOAIMessages_TextFileExtractedText) {
+  // Test that text files with extracted_text produce TextContentBlock with
+  // [File: filename] prefix, and text files without extracted_text are skipped.
+  auto history = CreateSampleChatHistory(1);
+
+  auto text_files =
+      CreateSampleUploadedFiles(2, mojom::UploadedFileType::kText);
+  // First text file has extracted text — should produce TextContentBlock
+  text_files[0]->extracted_text = "config_key=config_value";
+  text_files[0]->filename = "app.conf";
+  // Second text file has no extracted text — should be skipped
+  text_files[1]->filename = "failed.txt";
+
+  history[0]->uploaded_files = std::move(text_files);
+
+  PageContentsMap page_contents_map;
+  std::vector<OAIMessage> messages =
+      BuildOAIMessages(std::move(page_contents_map), history, nullptr, true,
+                       10000, [](std::string&) {});
+
+  ASSERT_EQ(messages.size(), 2u);
+  EXPECT_EQ(messages[0].role, "user");
+  // Content: text files intro + extracted text + prompt = 3 blocks
+  // (second file with no extracted_text is skipped)
+  ASSERT_EQ(messages[0].content.size(), 3u);
+  VerifyTextBlock(FROM_HERE, messages[0].content[0],
+                  "These text files are uploaded by the user");
+  VerifyTextBlock(FROM_HERE, messages[0].content[1],
+                  "[File: app.conf]\nconfig_key=config_value");
+  VerifyTextBlock(FROM_HERE, messages[0].content[2], "query0");
+}
+
+TEST_F(OAIMessageUtilsTest, BuildOAIMessages_TextFileDefaultFilename) {
+  // Test that text files with empty filename get default "uploaded.txt"
+  auto history = CreateSampleChatHistory(1);
+
+  auto text_files =
+      CreateSampleUploadedFiles(1, mojom::UploadedFileType::kText);
+  text_files[0]->extracted_text = "some content";
+  text_files[0]->filename.clear();
+
+  history[0]->uploaded_files = std::move(text_files);
+
+  PageContentsMap page_contents_map;
+  std::vector<OAIMessage> messages =
+      BuildOAIMessages(std::move(page_contents_map), history, nullptr, true,
+                       10000, [](std::string&) {});
+
+  ASSERT_EQ(messages.size(), 2u);
+  ASSERT_EQ(messages[0].content.size(), 3u);
+  VerifyTextBlock(FROM_HERE, messages[0].content[1],
+                  "[File: uploaded.txt]\nsome content");
+}
+
 TEST_F(OAIMessageUtilsTest, BuildOAIMessages_Memory_Excluded) {
   // Enable customization and set data
   prefs_.SetBoolean(prefs::kBraveAIChatUserCustomizationEnabled, true);
diff --git a/components/ai_chat/core/common/mojom/ai_chat.mojom b/components/ai_chat/core/common/mojom/ai_chat.mojom
index 88150211966..58633afdba3 100644
--- a/components/ai_chat/core/common/mojom/ai_chat.mojom
+++ b/components/ai_chat/core/common/mojom/ai_chat.mojom
@@ -218,6 +218,11 @@ interface AIChatUIHandler {
   ProcessPdfFile(array<uint8> file_data, string filename)
       => (UploadedFile? processed_file);
 
+  // Process text files: extract text via background renderer for encoding
+  // detection
+  ProcessTextFile(array<uint8> file_data, string filename)
+      => (UploadedFile? processed_file);
+
   // Get a plural string for the given key and count.
   GetPluralString(string key, int32 count) => (string plural_string);
 
diff --git a/components/ai_chat/core/common/mojom/common.mojom b/components/ai_chat/core/common/mojom/common.mojom
index 7783be9ea1e..e14e58a0dd5 100644
--- a/components/ai_chat/core/common/mojom/common.mojom
+++ b/components/ai_chat/core/common/mojom/common.mojom
@@ -113,6 +113,7 @@ enum UploadedFileType {
   kImage = 0,
   kScreenshot,
   kPdf,
+  kText,
 };
 
 // This does not cover more specific data that the Service owns, such as the
@@ -151,9 +152,9 @@ struct UploadedFile {
   uint32 filesize;
   array<uint8> data;
   UploadedFileType type;
-  // Text extracted from PDF files via the background PDF viewer and ScreenAI
-  // OCR pipeline. Null for non-PDF files or when extraction failed/timed out.
-  // When present, the engine layer sends this text instead of raw PDF bytes.
+  // Text extracted from PDF or text files via a background renderer. Null when
+  // extraction failed/timed out or not applicable. When present, the engine
+  // layer sends this text instead of raw file bytes.
   string? extracted_text;
 };
 
diff --git a/components/ai_chat/resources/common/conversation_history_utils.ts b/components/ai_chat/resources/common/conversation_history_utils.ts
index f1498271c01..a5ef4a34bc8 100644
--- a/components/ai_chat/resources/common/conversation_history_utils.ts
+++ b/components/ai_chat/resources/common/conversation_history_utils.ts
@@ -237,7 +237,8 @@ export const processUploadedFilesWithLimits = (
     const isImage =
       file.type === Mojom.UploadedFileType.kImage
       || file.type === Mojom.UploadedFileType.kScreenshot
-    const isDocument = file.type === Mojom.UploadedFileType.kPdf
+    const isPdf = file.type === Mojom.UploadedFileType.kPdf
+    const isText = file.type === Mojom.UploadedFileType.kText
     if (isImage) {
       const maxNewImages =
         Mojom.MAX_IMAGES - totalUploadedImages - currentPendingImages
@@ -245,7 +246,7 @@ export const processUploadedFilesWithLimits = (
         newFiles.push(file)
         currentImages++
       }
-    } else if (isDocument) {
+    } else if (isPdf) {
       const hasExtractedText = !!file.extractedText
       if (hasExtractedText) {
         // PDFs with extracted text bypass raw file limits
@@ -265,6 +266,9 @@ export const processUploadedFilesWithLimits = (
           currentRawDocuments++
         }
       }
+    } else if (isText) {
+      // Text files are not subject to document count/size limits
+      newFiles.push(file)
     }
   }
 
diff --git a/components/ai_chat/resources/page/api/ai_chat_api.ts b/components/ai_chat/resources/page/api/ai_chat_api.ts
index 8929b98cbff..c2b5f59b62e 100644
--- a/components/ai_chat/resources/page/api/ai_chat_api.ts
+++ b/components/ai_chat/resources/page/api/ai_chat_api.ts
@@ -104,6 +104,9 @@ export default function createAIChatApi(
         processPdfFile: {
           mutationResponse: (result) => result.processedFile,
         },
+        processTextFile: {
+          mutationResponse: (result) => result.processedFile,
+        },
         getPluralString: {
           response: (result) => result.pluralString,
         },
diff --git a/components/ai_chat/resources/page/api/mock_interfaces.ts b/components/ai_chat/resources/page/api/mock_interfaces.ts
index ef18b812fe5..f0c96bd5193 100644
--- a/components/ai_chat/resources/page/api/mock_interfaces.ts
+++ b/components/ai_chat/resources/page/api/mock_interfaces.ts
@@ -213,6 +213,16 @@ export function createMockUIHandler(
           extractedText: undefined,
         },
       }),
+    processTextFile: () =>
+      Promise.resolve({
+        processedFile: {
+          filename: '',
+          filesize: 0,
+          data: [],
+          type: Mojom.UploadedFileType.kText,
+          extractedText: undefined,
+        },
+      }),
     getPluralString: () => Promise.resolve({ pluralString: '' }),
     setChatUI: () => Promise.resolve({ isStandalone: false }),
 
diff --git a/components/ai_chat/resources/page/components/attachment_item/index.tsx b/components/ai_chat/resources/page/components/attachment_item/index.tsx
index dfb2fdb915a..a810df5dd25 100644
--- a/components/ai_chat/resources/page/components/attachment_item/index.tsx
+++ b/components/ai_chat/resources/page/components/attachment_item/index.tsx
@@ -175,6 +175,7 @@ function AttachmentUploadItem({
     file.type === Mojom.UploadedFileType.kImage
     || file.type === Mojom.UploadedFileType.kScreenshot
   const isPdf = file.type === Mojom.UploadedFileType.kPdf
+  const isText = file.type === Mojom.UploadedFileType.kText
   const isFileFullPageScreenshot = isFullPageScreenshot(file)
 
   const dataUrl = React.useMemo(() => {
@@ -208,7 +209,7 @@ function AttachmentUploadItem({
         className={className}
       />
     )
-  } else if (isPdf) {
+  } else if (isPdf || isText) {
     return (
       <AttachmentItem
         icon={<Icon name='file' />}
diff --git a/components/ai_chat/resources/page/components/drag_overlay/index.tsx b/components/ai_chat/resources/page/components/drag_overlay/index.tsx
index 5e38b70e164..9feb55ea59d 100644
--- a/components/ai_chat/resources/page/components/drag_overlay/index.tsx
+++ b/components/ai_chat/resources/page/components/drag_overlay/index.tsx
@@ -5,9 +5,10 @@
 
 import * as React from 'react'
 import { getLocale } from '$web-common/locale'
+import { showAlert } from '@brave/leo/react/alertCenter'
 import Icon from '@brave/leo/react/icon'
+import * as Mojom from '../../../common/mojom'
 import styles from './style.module.scss'
-import { isImageFile, isPdfFile } from '../../constants/file_types'
 import { useAIChat } from '../../state/ai_chat_context'
 import { useConversation } from '../../state/conversation_context'
 import { convertFileToUploadedFile } from '../../utils/file_utils'
@@ -21,25 +22,36 @@ export default function DragOverlay() {
     e.stopPropagation()
     clearDragState()
 
-    const files = Array.from(e.dataTransfer?.files || []).filter(
-      (file) => isImageFile(file) || isPdfFile(file),
-    )
+    const files = Array.from(e.dataTransfer?.files || [])
 
     if (files.length === 0) {
       return
     }
 
     try {
-      const uploadedFiles = await Promise.all(
+      const results = await Promise.all(
         files.map((file) =>
           convertFileToUploadedFile(
             file,
             aiChat.processImageFile,
             aiChat.processPdfFile,
+            aiChat.processTextFile,
           ),
         ),
       )
-      attachImages(uploadedFiles)
+      const uploadedFiles = results.filter(
+        (f): f is Mojom.UploadedFile => f !== null,
+      )
+      if (uploadedFiles.length > 0) {
+        attachImages(uploadedFiles)
+      }
+      if (uploadedFiles.length < files.length) {
+        showAlert({
+          type: 'error',
+          content: getLocale(S.CHAT_UI_FILE_UPLOAD_ERROR),
+          actions: [],
+        })
+      }
     } catch (error) {
       // Silently fail - error will be handled by the upload system
     }
diff --git a/components/ai_chat/resources/page/components/input_box/index.test.tsx b/components/ai_chat/resources/page/components/input_box/index.test.tsx
index 4769e35665d..5d27592e585 100644
--- a/components/ai_chat/resources/page/components/input_box/index.test.tsx
+++ b/components/ai_chat/resources/page/components/input_box/index.test.tsx
@@ -436,16 +436,23 @@ describe('input box', () => {
       jest.clearAllMocks()
 
       mockConvertFileToUploadedFile.mockImplementation((file: File) => {
+        const mimeType = file.type.toLowerCase()
+        let type = UploadedFileType.kText
+        if (mimeType.startsWith('image/')) {
+          type = UploadedFileType.kImage
+        } else if (mimeType === 'application/pdf') {
+          type = UploadedFileType.kPdf
+        }
         return Promise.resolve({
           filename: file.name,
           filesize: file.size,
-          data: Array.from(new Uint8Array(8)), // Mock data array
-          type: UploadedFileType.kImage,
+          data: Array.from(new Uint8Array(8)),
+          type,
         })
       })
     })
 
-    it('filters image files and calls attachImages on paste', async () => {
+    it('accepts all file types on paste', async () => {
       const mockAttachImages = jest.fn()
       const { container } = await renderInputBox(
         <MockContext>
@@ -488,6 +495,12 @@ describe('input box', () => {
             data: expect.any(Array),
             type: UploadedFileType.kImage,
           }),
+          expect.objectContaining({
+            filename: 'test.txt',
+            filesize: 1024,
+            data: expect.any(Array),
+            type: UploadedFileType.kText,
+          }),
         ])
       })
     })
diff --git a/components/ai_chat/resources/page/components/input_box/index.tsx b/components/ai_chat/resources/page/components/input_box/index.tsx
index 464ec149cd6..0cb6df566cb 100644
--- a/components/ai_chat/resources/page/components/input_box/index.tsx
+++ b/components/ai_chat/resources/page/components/input_box/index.tsx
@@ -3,6 +3,7 @@
  * License, v. 2.0. If a copy of the MPL was not distributed with this file,
  * You can obtain one at https://mozilla.org/MPL/2.0/. */
 
+import { showAlert } from '@brave/leo/react/alertCenter'
 import Icon from '@brave/leo/react/icon'
 import Button from '@brave/leo/react/button'
 import Tooltip from '@brave/leo/react/tooltip'
@@ -23,7 +24,6 @@ import {
 } from '../attachment_item'
 import { ModelSelector } from '../model_selector'
 import usePromise from '$web-common/usePromise'
-import { isImageFile } from '../../constants/file_types'
 import { convertFileToUploadedFile } from '../../utils/file_utils'
 import { isFullPageScreenshot } from '../../../common/conversation_history_utils'
 import Editable from './editable'
@@ -76,6 +76,7 @@ type Props = Pick<
     | 'getPluralString'
     | 'processImageFile'
     | 'processPdfFile'
+    | 'processTextFile'
     | 'openAIChatAgentProfile'
     | 'skills'
   >
@@ -219,26 +220,39 @@ function InputBox(props: InputBoxProps) {
       return
     }
 
-    const files = Array.from(clipboardData.files).filter(isImageFile)
+    const files = Array.from(clipboardData.files)
 
     if (files.length === 0) {
       return
     }
 
-    // Prevent the default paste behavior for images
+    // Prevent the default paste behavior for files
     e.preventDefault()
 
     try {
-      const uploadedFiles = await Promise.all(
+      const results = await Promise.all(
         files.map((file) =>
           convertFileToUploadedFile(
             file,
             props.context.processImageFile,
             props.context.processPdfFile,
+            props.context.processTextFile,
           ),
         ),
       )
-      props.context.attachImages(uploadedFiles)
+      const uploadedFiles = results.filter(
+        (f): f is Mojom.UploadedFile => f !== null,
+      )
+      if (uploadedFiles.length > 0) {
+        props.context.attachImages(uploadedFiles)
+      }
+      if (uploadedFiles.length < files.length) {
+        showAlert({
+          type: 'error',
+          content: getLocale(S.CHAT_UI_FILE_UPLOAD_ERROR),
+          actions: [],
+        })
+      }
     } catch (error) {
       // Silently fail - error will be handled by the upload system
     }
diff --git a/components/ai_chat/resources/page/state/ai_chat_context.tsx b/components/ai_chat/resources/page/state/ai_chat_context.tsx
index 8e82336a56d..82358067443 100644
--- a/components/ai_chat/resources/page/state/ai_chat_context.tsx
+++ b/components/ai_chat/resources/page/state/ai_chat_context.tsx
@@ -152,6 +152,7 @@ export default function useProvideAIChatContext(props: AIChatContextProps) {
     // and we can do that via monitoring the mutation in the provided hook.
     processImageFile: api.processImageFile,
     processPdfFile: api.processPdfFile,
+    processTextFile: api.processTextFile,
 
     /**
      * @deprecated use api.uiHandler.openAIChatAgentProfile directly instead
diff --git a/components/ai_chat/resources/page/state/conversation_context.tsx b/components/ai_chat/resources/page/state/conversation_context.tsx
index 5ccabb5b0c4..ed9c306c134 100644
--- a/components/ai_chat/resources/page/state/conversation_context.tsx
+++ b/components/ai_chat/resources/page/state/conversation_context.tsx
@@ -4,7 +4,9 @@
 // You can obtain one at https://mozilla.org/MPL/2.0/.
 
 import * as React from 'react'
+import { showAlert } from '@brave/leo/react/alertCenter'
 import generateReactContext from '$web-common/api/react_api'
+import { getLocale } from '$web-common/locale'
 import { Url } from 'gen/url/mojom/url.mojom.m.js'
 import { IGNORE_EXTERNAL_LINK_WARNING_KEY } from '../../common/constants'
 import {
@@ -391,6 +393,24 @@ export function useProvideConversationContext(props: ConversationContextProps) {
   }
 
   const processUploadedFiles = async (files: Mojom.UploadedFile[]) => {
+    // Filter out text files where extraction failed (no extracted text).
+    const validFiles = files.filter(
+      (f) =>
+        f.type !== Mojom.UploadedFileType.kText
+        || (f.extractedText !== undefined && f.extractedText !== null),
+    )
+    // Show error when some files were dropped: either text extraction
+    // failed, or unsupported files were included (the backend returns
+    // empty stubs for unsupported types like zip so they are filtered
+    // out here, while cancellation returns null and skips this path).
+    if (validFiles.length < files.length) {
+      showAlert({
+        type: 'error',
+        content: getLocale(S.CHAT_UI_FILE_UPLOAD_ERROR),
+        actions: [],
+      })
+    }
+
     // After mutation, any returned promise will be awaited before settling.
     // This won't re-fetch the conversation history, just get the latest
     // version if it's not invalidated.
@@ -399,7 +419,7 @@ export function useProvideConversationContext(props: ConversationContextProps) {
     // data.
     setPendingMessageFiles((pendingMessageFiles) => {
       const newFiles = processUploadedFilesWithLimits(
-        files,
+        validFiles,
         conversationHistory,
         pendingMessageFiles,
       )
@@ -433,7 +453,8 @@ export function useProvideConversationContext(props: ConversationContextProps) {
   ) => {
     uploadFileMutation.mutate(args, {
       onSuccess: async (uploadedFiles, [useMediaCapture]) => {
-        // Reset event state, avoid us having to make a useState<bool> for this
+        // Reset event state, avoid us having to make a useState<bool>
+        // for this
         aiChat.api.resetOnUploadFilesSelected()
         if (uploadedFiles) {
           return processUploadedFiles(uploadedFiles)
diff --git a/components/ai_chat/resources/page/utils/file_utils.test.ts b/components/ai_chat/resources/page/utils/file_utils.test.ts
index e90fe9f8a66..f77a9296108 100644
--- a/components/ai_chat/resources/page/utils/file_utils.test.ts
+++ b/components/ai_chat/resources/page/utils/file_utils.test.ts
@@ -7,7 +7,6 @@ import {
   convertFileToUploadedFile,
   FileReadError,
   ImageProcessingError,
-  UnsupportedFileTypeError,
 } from './file_utils'
 import * as Mojom from '../../common/mojom'
 import type { AIChatContext } from '../state/ai_chat_context'
@@ -241,12 +240,18 @@ describe('convertFileToUploadedFile', () => {
       )
     })
 
-    it('throws UnsupportedFileTypeError for unknown file types', async () => {
+    it('treats text files as kText via processTextFile', async () => {
       const file = createMockFile('test.txt', 'text/plain')
       const mockArrayBuffer = new ArrayBuffer(8)
+      const expectedData = Array.from(new Uint8Array(mockArrayBuffer))
+      const mockProcessTextFile = jest.fn().mockResolvedValue({
+        filename: 'test.txt',
+        filesize: file.size,
+        data: expectedData,
+        type: Mojom.UploadedFileType.kText,
+        extractedText: 'extracted content',
+      })
 
-      // Override readAsArrayBuffer to trigger success
-      // (FileReader will work, but file type check will fail)
       mockFileReader.readAsArrayBuffer.mockImplementation(() => {
         process.nextTick(() => {
           if (mockFileReader.onload) {
@@ -255,16 +260,45 @@ describe('convertFileToUploadedFile', () => {
         })
       })
 
-      await expect(
-        convertFileToUploadedFile(file, mockProcessImageFile),
-      ).rejects.toThrow(UnsupportedFileTypeError)
-      await expect(
-        convertFileToUploadedFile(file, mockProcessImageFile),
-      ).rejects.toThrow(
-        'Unsupported file type: text/plain. Only images and PDF files are '
-          + 'supported.',
+      const result = await convertFileToUploadedFile(
+        file,
+        mockProcessImageFile,
+        undefined,
+        mockProcessTextFile,
       )
+
       expect(mockProcessImageFile).not.toHaveBeenCalled()
+      expect(mockProcessTextFile).toHaveBeenCalled()
+      expect(result).toEqual(
+        expect.objectContaining({
+          filename: 'test.txt',
+          type: Mojom.UploadedFileType.kText,
+          extractedText: 'extracted content',
+        }),
+      )
+    })
+
+    it('returns null for text files when extraction fails', async () => {
+      const file = createMockFile('binary.dat', 'application/octet-stream')
+      const mockArrayBuffer = new ArrayBuffer(8)
+      const mockProcessTextFile = jest.fn().mockResolvedValue(null)
+
+      mockFileReader.readAsArrayBuffer.mockImplementation(() => {
+        process.nextTick(() => {
+          if (mockFileReader.onload) {
+            mockFileReader.onload({ target: { result: mockArrayBuffer } })
+          }
+        })
+      })
+
+      const result = await convertFileToUploadedFile(
+        file,
+        mockProcessImageFile,
+        undefined,
+        mockProcessTextFile,
+      )
+
+      expect(result).toBeNull()
     })
   })
 
diff --git a/components/ai_chat/resources/page/utils/file_utils.ts b/components/ai_chat/resources/page/utils/file_utils.ts
index 0f1b2771162..577a1110c4d 100644
--- a/components/ai_chat/resources/page/utils/file_utils.ts
+++ b/components/ai_chat/resources/page/utils/file_utils.ts
@@ -27,19 +27,13 @@ export class ImageProcessingError extends Error {
   }
 }
 
-export class UnsupportedFileTypeError extends Error {
-  constructor(message: string) {
-    super(message)
-    this.name = 'UnsupportedFileTypeError'
-  }
-}
-
 // Utility function to convert File objects to UploadedFile format
 export const convertFileToUploadedFile = async (
   file: File,
   processImageFile: AIChatContext['processImageFile'],
   processPdfFile?: AIChatContext['processPdfFile'],
-): Promise<Mojom.UploadedFile> => {
+  processTextFile?: AIChatContext['processTextFile'],
+): Promise<Mojom.UploadedFile | null> => {
   const reader = new FileReader()
   const arrayBuffer = await new Promise<ArrayBuffer>((resolve, reject) => {
     reader.onload = (e) => {
@@ -70,41 +64,41 @@ export const convertFileToUploadedFile = async (
       }
     }
     // Fallback: return raw PDF data without extracted text
-    const uploadedFile: Mojom.UploadedFile = {
+    return {
       filename: file.name,
       filesize: file.size,
       data: Array.from(uint8Array),
       type: Mojom.UploadedFileType.kPdf,
       extractedText: undefined,
     }
-    return uploadedFile
   } else if (mimeType.startsWith('image/')) {
     // Use backend processing for images via mojo call
-    try {
-      const processedFile = await processImageFile([
+    const processedFile = await processImageFile([
+      Array.from(uint8Array),
+      file.name,
+    ])
+
+    if (!processedFile) {
+      throw new ImageProcessingError(
+        'Failed to process image file: Backend returned no result',
+      )
+    }
+
+    return processedFile
+  } else {
+    // Everything else is treated as a text file. The renderer handles
+    // MIME sniffing and will render the file as text if possible.
+    // If extraction fails (e.g. binary file), null is returned and
+    // the file is not attached.
+    if (processTextFile) {
+      const processedFile = await processTextFile([
         Array.from(uint8Array),
         file.name,
       ])
-
-      if (!processedFile) {
-        throw new ImageProcessingError(
-          'Failed to process image file: Backend returned no result',
-        )
+      if (processedFile) {
+        return processedFile
       }
-
-      return processedFile
-    } catch (error) {
-      if (error instanceof ImageProcessingError) {
-        throw error
-      }
-
-      // Re-throw any other errors as-is
-      throw error
     }
-  } else {
-    throw new UnsupportedFileTypeError(
-      `Unsupported file type: ${file.type}. Only images and PDF files `
-        + `are supported.`,
-    )
+    return null
   }
 }
diff --git a/components/resources/ai_chat_ui_strings.grdp b/components/resources/ai_chat_ui_strings.grdp
index 00f57aea4c2..661d242dce2 100644
--- a/components/resources/ai_chat_ui_strings.grdp
+++ b/components/resources/ai_chat_ui_strings.grdp
@@ -45,6 +45,9 @@
   <message name="IDS_CHAT_UI_DROP_FILES_DESCRIPTION" desc="Descriptive text shown when dragging files over the page" formatter_data="webui=AiChat">
     Drop any files here to add them to the conversation
   </message>
+  <message name="IDS_CHAT_UI_FILE_UPLOAD_ERROR" desc="Alert shown when one or more files could not be read as text and were not attached" formatter_data="webui=AiChat">
+    Some files could not be read and were not attached
+  </message>
   <message name="IDS_CHAT_UI_SUGGEST_QUESTIONS_LABEL" desc="Button label to ask the AI Chat to suggest some related questions" formatter_data="webui=AiChat">
     Suggest questions…
   </message>
diff --git a/ios/browser/ui/webui/ai_chat/ai_chat_ui_page_handler.h b/ios/browser/ui/webui/ai_chat/ai_chat_ui_page_handler.h
index 269b6a04ffe..238f92e52ad 100644
--- a/ios/browser/ui/webui/ai_chat/ai_chat_ui_page_handler.h
+++ b/ios/browser/ui/webui/ai_chat/ai_chat_ui_page_handler.h
@@ -66,6 +66,9 @@ class AIChatUIPageHandler : public mojom::AIChatUIHandler,
   void ProcessPdfFile(const std::vector<uint8_t>& file_data,
                       const std::string& filename,
                       ProcessPdfFileCallback callback) override;
+  void ProcessTextFile(const std::vector<uint8_t>& file_data,
+                       const std::string& filename,
+                       ProcessTextFileCallback callback) override;
   void UploadFile(bool use_media_capture, UploadFileCallback callback) override;
   void GetPluralString(const std::string& key,
                        int32_t count,
diff --git a/ios/browser/ui/webui/ai_chat/ai_chat_ui_page_handler.mm b/ios/browser/ui/webui/ai_chat/ai_chat_ui_page_handler.mm
index 356eb50eeef..c94f0cbb05c 100644
--- a/ios/browser/ui/webui/ai_chat/ai_chat_ui_page_handler.mm
+++ b/ios/browser/ui/webui/ai_chat/ai_chat_ui_page_handler.mm
@@ -225,6 +225,17 @@ void AIChatUIPageHandler::ProcessPdfFile(const std::vector<uint8_t>& file_data,
   std::move(callback).Run(std::move(uploaded_file));
 }
 
+void AIChatUIPageHandler::ProcessTextFile(const std::vector<uint8_t>& file_data,
+                                          const std::string& filename,
+                                          ProcessTextFileCallback callback) {
+  // iOS does not support background text extraction.
+  // Return the raw data without extracted text.
+  auto uploaded_file = ai_chat::mojom::UploadedFile::New(
+      filename, file_data.size(), file_data,
+      ai_chat::mojom::UploadedFileType::kText, std::nullopt);
+  std::move(callback).Run(std::move(uploaded_file));
+}
+
 void AIChatUIPageHandler::UploadFile(bool use_media_capture,
                                      UploadFileCallback callback) {
   id<AIChatUIHandlerBridge> bridge =
diff --git a/test/data/leo/dummy.html b/test/data/leo/dummy.html
new file mode 100644
index 00000000000..861b69885fb
--- /dev/null
+++ b/test/data/leo/dummy.html
@@ -0,0 +1,9 @@
+<!DOCTYPE html>
+<html>
+<head><title>Test</title></head>
+<body>
+<p>Hello from an HTML file.</p>
+<img src="https://example.com/track.png">
+<script>document.title = "executed";</script>
+</body>
+</html>
diff --git a/test/data/leo/dummy.txt b/test/data/leo/dummy.txt
new file mode 100644
index 00000000000..9ba06051f86
--- /dev/null
+++ b/test/data/leo/dummy.txt
@@ -0,0 +1,2 @@
+Hello from a text file.
+This is line two.
\ No newline at end of file