Browser Tools

Browser tools connect your web research, AI conversations, and online content to your local ThinkForge projects — without leaving your browser.

When the Chrome extension is connected to ThinkForge Desktop, you can capture content from any webpage and send it directly into a project. Everything is saved as local files and indexed into project memory.

What you can capture

  • Full web pages — save an entire page's content into your project as a Markdown file
  • Text selections — highlight text on any page and send just that selection to a project
  • AI conversations — capture full chat threads from ChatGPT, Claude, Gemini, and other AI platforms with source metadata preserved
  • Images — save images from web pages into your project folder
  • PDFs — upload PDF files into a project via the extension

How capturing works

  1. Browse the web as normal with the ThinkForge extension installed and connected.
  2. When you find content worth saving, use the extension's quick command menu, right-click context menu, or the side panel controls.
  3. Choose the target project (or use the currently active one).
  4. The content is sent to ThinkForge Desktop over your local connection.
  5. ThinkForge saves it as a file in your project folder (typically under _captured) and indexes it into project memory.

Captured content includes metadata about where it came from — the source URL, page title, capture date, and (for AI chats) the platform name.

AI chat capture

ThinkForge has built-in support for capturing conversations from major AI platforms. The extension detects when you are on a supported chat interface and offers conversation-specific capture options.

Supported platforms include ChatGPT, Claude, Gemini, and others. When you capture a chat:

  • The full conversation thread is extracted (questions and responses)
  • Source metadata (platform, chat URL, title) is preserved in the saved file
  • The content is saved as a Markdown file with optional YAML frontmatter
  • Each question-and-answer pair is indexed as a searchable chunk in project memory

Desktop-initiated capture

Capture also works in the other direction. From ThinkForge Desktop you can trigger a browser capture through the command menu or from the Project Navigator. This tells the extension to grab content from the current browser tab and send it back to the desktop app — useful when you are working in ThinkForge and spot something in your browser you want to save without switching focus.

Where captured content goes

By default, captures land in the _captured folder inside your active project. You can also choose a specific folder when capturing. Files are saved as Markdown (.md) for text content, or in their original format for images and PDFs.

Once saved, captured content appears in the Project Navigator, is searchable through project memory, and can be referenced by agents and in chat context.

All local. Captured content travels directly from your browser to ThinkForge Desktop over localhost. It is saved as files on your computer and indexed locally. Nothing is uploaded to external servers as part of the capture process.

Related