BrowsePal

BrowsePal is a free privacy-first AI copilot that lives in Chrome’s side panel. • Chat with any web page, summarize articles, extract key points from PDFs, pull insights from YouTube videos, and let the assistant control your browser when you want it to. • Bring your own API key, choose your preferred model, and keep full control of your data. • BrowsePal supports Anthropic Claude, OpenAI GPT, Google Gemini, xAI Grok, OpenRouter, and OpenAI-compatible endpoints such as Ollama, LM Studio, and vLLM. • New models appear in the picker as soon as your provider lists them, with no extension update required. ◆ Why BrowsePal? • True agentic experience: Using Skills, MCP servers, and Browser Use, BrowsePal can understand your context, access external tools, and take user-approved actions directly in your browser. • Any model, any provider: Pay your provider directly. No proxy, no markup, no prompt logging. • Privacy by design: Your API keys are encrypted on-device with AES-GCM 256 and never leave your browser. • Optional zero-knowledge sync: Sign in with a magic link, set a passphrase, and sync your settings, MCP servers, and chat history across devices with end-to-end encryption. • Zero telemetry: No analytics, no tracking, no ads. ◆ Chat With The Web Attach rich context to any message: • Current page content, cleaned with Mozilla Readability • Text selection from the active tab • Open tabs or content from another tab • Viewport, full-page, or region screenshots • Local files including TXT, Markdown, PDF, and images • YouTube transcripts with timestamps and metadata • Images pasted directly from your clipboard BrowsePal is great for researching articles, comparing documents across tabs, drafting replies, summarizing long PDFs, and turning YouTube videos into clear bullet points. ◆ Agentic Browser Control Turn on Tool Use and your model can control the active tab with built-in browser tools: • Snapshot • Click • Type • Select • Press key • Scroll • Navigate • Wait for element • Take screenshot • Manage tabs BrowsePal can help complete multi-step tasks such as filling forms, searching websites, comparing pages, or walking through checkout flows with accessible DOM grounding. ◆ Custom Skills and Instructions • Teach BrowsePal your workflow. • Add up to five agentskills.io-format skills using SKILL.md files. • Use skills for tasks like code review, security audits, writing release notes, or following your own internal process. • Add custom instructions to shape the assistant’s tone, output format, or behavior across every chat. ◆ MCP Servers, Native • Connect any Model Context Protocol server over Streamable HTTP or legacy SSE. • Add custom auth headers. • Toggle servers per chat. • Expose MCP tools to any supported model. • Use your own MCP server for code, tickets, internal docs, or private APIs without a separate plugin layer. ◆ BYOK Done Right • BYOK means Bring Your Own Key. • You are billed directly by your AI provider for the tokens you use. • No proxy. • No subscription lock-in. • No hidden quota system. • Paste your API key once, BrowsePal encrypts it locally, and model calls go directly to your provider. ◆ Customize Everything • Light, dark, or system theme • IDE and Studio layout variations • Five accent colors: violet, mint, sky, orange, and pink • Chat history • Auto-generated chat titles • Keyboard-friendly composer • Paste-to-attach workflow • Clean, slash-free interface ◆ Perfect For • Developers reading API docs • Researchers comparing papers • Students summarizing lectures • Writers drafting from sources • PMs taking notes across tabs • Power users who want an AI assistant that respects their keys and privacy ◆ Need Help? • Visit https://www.browsepal.ai for more info.