tab-agent

Monorepo for Local tab agent: a VS Code coding assistant that uses Ollama locally and ChatGPT, Gemini, or Claude in the browser via a Chrome bridge.

Release 1.1.0 (highlights)

• Groq API integration — Fast, low-latency code generation using Groq. API keys are stored securely using VS Code's SecretStorage.
• Collaborative Swarm Groups — Pool CPU/GPU computing power collaboratively with friends and colleagues to run large models as a single distributed agent.
• LLM Server Auto-Discovery — Probes localhost and local network for running Ollama/OpenAI servers, measures latency, and configures endpoints with one click.
• Incremental Vector Indexing & Search — Automatically indexes codebase files on save to build a semantic search database for smarter context selection.
• Auto-Fix Diagnostics — Automatically trigger code modifications to solve workspace compiler warnings/errors.
• Hybrid Routing — Balance local Ollama and remote backends based on task complexity.
• Documentation Generation — A new command (aiagent.generateDocs) to automatically document files.

Full extension notes: vscode-extension/README.md.

Workflow

Swarm Groups Dashboard

Exposing the Swarm Broker Across Networks

By default, the Swarm Broker (located in vscode-extension/broker) runs on ws://localhost:3000. To collaborate with friends on different physical networks, the broker must be publicly accessible. You can do this in three ways:

• Option A: Deploy to the Cloud (Recommended) Deploy the Node.js/TypeScript broker project to free or low-cost hosting services like Render, Railway, or Fly.io, or on a VPS (AWS, DigitalOcean). Ensure standard setup runs npm install and npm start. Connect your extension using secure WebSocket format: wss://your-broker-app.onrender.com.
• Option B: Local Tunneling (Easiest for quick sessions and free ) Use tunneling tools to expose your local port 3000 to the web:
  • ngrok: Run ngrok http 3000 to get a public URL (e.g. https://xxxx.ngrok-free.app), and connect extensions via its WebSocket secure mapping wss://xxxx.ngrok-free.app.
  • Cloudflare Tunnels: Point a tunnel to localhost:3000 and use your mapped hostname.
• Option C: Mesh VPNs (Tailscale / ZeroTier) If you and your friends are part of the same Tailscale network, you can run the broker locally and connect directly using the host machine's private Tailscale IP (e.g. ws://100.x.y.z:3000).

Part	Folder	Description
VS Code extension	vscode-extension/	Sidebar agent, Ollama, WebSocket bridge, inline review UI
Chrome extension	chrome-extension/	Injects prompts into the AI tab (ChatGPT / Gemini / Claude) and returns replies

• Repository: github.com/moego0/tab-agent

Quick start

1. Ollama — install and run ollama serve, pull a coding model.
2. VS Code extension — see vscode-extension/README.md: npm install && npm run build, then F5 or package with vsce.
3. Chrome — follow Install Local tab bridge below, then keep a ChatGPT, Gemini, or Claude tab open (match the provider selected in the sidebar).

Install Local tab bridge (Chrome)

1. Click the puzzle icon (Extensions) in the Chrome toolbar → Manage extensions.
2. Turn on Developer mode (top right), then click Load unpacked.
3. Choose the chrome-extension folder from this repo (the one that contains manifest.json) → Select Folder.

Step 1 — Manage extensions

Step 2 — Load unpacked

Step 3 — Select the chrome-extension folder

Contribution

Contributions are very welcome. The goal of this project is to turn your subscription to any LLM into a coding agent, helping users reduce API costs while still getting powerful AI-assisted development. It also works without a paid plan; free tiers may limit context depending on the provider.

More detail: chrome-extension/README.md.

License

MIT — see vscode-extension/package.json.