Freekit

Free AI coding agent for VS Code. Use your existing web AI accounts (DeepSeek, ChatGPT, Claude, Gemini, Mistral, Qwen) or a local Ollama model. No API keys, no subscriptions, no costs.

Screenshots

Why Freekit?

Unlike other AI coding agents, Freekit works with your existing web AI accounts:

Perfect for:

• Students and hobbyists without API budgets
• Developers in emerging markets (cost-sensitive)
• Anyone who already has a ChatGPT Plus / Claude Pro subscription
• Privacy-conscious users (local Ollama option)

Features

• 🎯 Six web AI providers + Ollama — DeepSeek, ChatGPT, Claude, Gemini, Mistral, Qwen + local Ollama
• 🔒 Stealth mode — human-like typing (real keydown/keyup), anti-detection
• 🤖 Agentic loop — up to 40 iterations, 12 tools (read/write/edit/search/run/git)
• 📝 Native diff review — Accept/Reject each file change in VS Code's diff viewer
• ⟲ Git checkpoints — restore any prompt with one click
• ✂️ Fork & edit conversations — ChatGPT-style conversation branching
• 🔊 TTS + 🎤 voice input — offline Whisper transcription
• 📚 Semantic code search — find code by meaning, not just text
• 🔌 MCP servers — Model Context Protocol support
• 🛡️ Hardware-aware Ollama — 51-model catalog, tiers T0-T6, auto-recommendation
• 🧪 Auto-verify & auto-repair — every file write is checked (tsc --noEmit / astro check / build); failures go back to the AI (up to 3 attempts) and are rolled back if they can't be fixed
• 🖥️ Dev servers in a visible terminal — npm run dev, vite, nodemon, … start in a dedicated VS Code terminal, and the AI gets the live URL immediately
• 🔍 Verbose mode — collapsible Thinking / Executing / Result / Decision cards for full transparency
• 📎 Attachments & auto-context — files, folders, images and binaries, plus project structure sent with every message

Quick start

1. Install Freekit from the VS Code Marketplace
2. Open the sidebar → click the F icon
3. Click the model chip at the bottom → select a provider (e.g., DeepSeek)
4. A Chrome window opens → log in once (your session is saved)
5. Start chatting — try: Write a Python function for factorial

First time? VS Code may show a blue "Restricted Mode" banner. Click Trust to enable full functionality (Freekit writes files and runs commands — this is required).

Supported providers

Web (browser automation — no API key needed)

• DeepSeek (Chat V3 / DeepThink R1)
• ChatGPT (GPT-4o / GPT-4.1 / o3)
• Claude (Sonnet / Opus / Haiku)
• Gemini (2.5 Pro / 2.5 Flash)
• Mistral (Vibe)
• Qwen

Local

• Ollama — 51-model catalog, hardware tiers T0-T6, auto-recommendation based on your VRAM and RAM

How it works

Freekit connects to your existing web AI sessions via browser automation:

1. Launches an isolated Chrome profile (separate from your personal Chrome)
2. You log in once to each provider (session is saved)
3. Freekit types messages with human-like keystrokes (real keydown/keyup, random delays)
4. It reads the AI response and renders it in the VS Code sidebar
5. File changes appear as native VS Code diffs — you Accept or Reject each one

Everything is local. No middle server. No data sent to third parties.

Privacy

• ✅ Zero telemetry — no analytics, no tracking
• ✅ Isolated browser — separate Chrome profile, doesn't touch your personal browser
• ✅ Local by default — Chrome profile, history and settings stay on your disk
• ⚠️ Browser automation — messages go through the web AI (same as if you typed them manually)
• ⚙️ Self-reporting (opt-in, on by default) — anonymous selector break reports to help fix issues. No code, files or messages sent. Disable: freekit.reporting.enabled
• 🔒 Ollama option — fully offline, nothing leaves your machine

Trust / Restricted Mode

Freekit writes files, runs terminal commands and controls a browser — it needs a trusted workspace.

On first launch, VS Code shows a blue "Restricted Mode" banner. Click Trust to enable full functionality.

In Restricted Mode (no trust):

• ✅ Chat UI works
• ✅ Read-only tools (read_file, search_files, etc.)
• ❌ File writes, terminal commands, git mutations, MCP tools

You can also run Freekit: Trust This Workspace from the Command Palette.

Additional safeguards:

• Writes open a native VS Code diff for review — accept or reject from the in-chat card or the VS Code notification; shell commands and git operations require approval cards unless ⚡ auto-approve is enabled.
• Anti-spam limits: max 3 writes per file and 15 write operations per message.
• The webview runs with a strict Content Security Policy.

Commands

Settings

Advanced

Semantic code search

Freekit can find code by meaning instead of exact text — useful for questions like "where do we retry a failed upload?" when you don't know the identifier to grep for.

1. Run Freekit: Index Workspace once (needs Ollama running; the embedding model is pulled with ollama pull nomic-embed-text).
2. Ask the AI something conceptual. It calls search_semantic(query) and gets back the best-matching chunks with path:startLine-endLine, a similarity score and a snippet — then it can read_file the ones that matter.

How it works:

• Local & private — embeddings are computed by your own Ollama server (freekit.ollamaUrl); the vector store is a single JSON file under <globalStorage>/semantic-index/ (keyed per workspace). No new npm dependencies, no cloud.
• Incremental — each file is fingerprinted with a SHA-1 of its content; unchanged files keep their existing vectors, so re-indexing after an edit only re-embeds what changed. Set freekit.semanticIndex.onSave to keep the index fresh automatically (debounced 1.5 s).
• Smart exclusions — node_modules, .git, out, dist, build, .next, coverage, virtualenvs, target, caches, lockfiles, source maps, minified and binary files are never indexed; extend the list with freekit.semanticIndex.exclude.
• Chunking with overlap — files are split into 60-line windows with a 12-line overlap so a function that straddles a boundary is still found; oversized/minified files are skipped and recorded in the status.
• Graceful in the agent loop — if there is no index yet, search_semantic returns a clear message telling the AI to fall back to search_files (or you to run the index command), so nothing breaks.

Remote selector updates

Selector repairs can reach every client without shipping a new extension version:

1. Keep the same JSON shape as src/selectors.json and bump version (updated / changelog are optional but recommended).
2. Save it as a file named selectors.json in a public GitHub Gist.
3. Set freekit.selectorsUrl to the Gist raw URL, e.g. https://gist.githubusercontent.com/<user>/<id>/raw/selectors.json.

Freekit then checks on startup — rate-limited to once every 24 h — and on demand via Freekit: Update Selectors, applying the config only when its version is newer than the active one. The update is validated, merged over the bundled config, cached locally, and local auto-repairs for the slots it touches are replaced by the curated fix. Any failure (invalid JSON, HTTP error, timeout) leaves the current config in place.

Reporting server (self-maintaining)

Freekit can feed a central reporting server (https://api.builderweb.app by default) so a site redesign discovered on one machine turns into a fix for everybody — without waiting for an extension release.

• Register once — on first start the extension creates a random installation id (globalStorage) and exchanges it for an apiKey (POST /api/v1/extensions/register). The call is idempotent: the key is cached and re-used until you run Freekit: Reset Reporting Registration.
• Health check every 6 h (freekit.reporting.healthCheckIntervalHours) — for every provider that already has a tab open in the Freekit Chrome profile, the input and newChat selectors are verified. Chrome is never started just for this. Broken slots are repaired locally first (fingerprint healer → AI finder); only if that also fails is a report sent (POST /api/v1/reports) with the domain, the failing selector, the failure type (not_found / hidden) and the page URL.
• Selector updates every 24 h (freekit.reporting.selectorsIntervalHours) — the extension asks GET /api/v1/selectors?since=<revision>, applies anything new (validated by the same fragile/blacklist rules as local repairs) and remembers the revision, so a 304 is the normal, cheap answer.
• Privacy — nothing from your code, prompts or files is ever sent. The only optional payload is the cleaned DOM snapshot (scripts, styles, media and input values are stripped, truncated) and it is off by default (freekit.reporting.shareDomSnapshot). Set freekit.reporting.enabled to false to opt out completely.

Use Freekit: Reporting Status for the current state (registration, revision, last check), Freekit: Run Selector Health Check to run a check immediately (e.g. right after a site redesign) and Freekit: Reset Reporting Registration to start over with a new identity.

MCP servers

Freekit ships a Model Context Protocol client: MCP servers give the agent extra tools (databases, browsers, APIs, file systems, …) beyond the built-in ones. Configure them once — they are launched with VS Code and their tools are exposed to the AI as mcp_<server>_<tool>.

1. Add servers to .vscode/mcp.json (or to the freekit.mcpServers setting):

{
  "servers": {
    "filesystem": {
      "command": "npx",
      "args": ["-y", "@modelcontextprotocol/server-filesystem", "."]
    }
  }
}

2. Servers start with VS Code (freekit.mcpEnabled, default on) and config edits are hot-reloaded. Use Freekit: MCP Servers to check status, restart/stop a server, or list its tools.

Every MCP call shows an approval card (server, tool and arguments) before it runs — ⚡ auto-approve skips it like any other tool. Only the stdio transport is supported. A crashed or missing server is reported in the chat and in Freekit: Diagnostics.

Troubleshooting

"Could not find the input box"

→ Log in to the provider. Click Show Chrome in the ⋯ menu, sign in, then retry.

"Restricted Mode — some features disabled"

→ Click Trust in the blue banner at the top of VS Code.

"Ollama not running"

→ Install Ollama from ollama.com and run ollama serve.

Browser automation issues

→ Click Freekit: Reset Repaired Selectors and try again.

Requirements

• VS Code 1.90+
• Chrome or Edge (installed)
• Ollama (optional, for local models)
• Windows / macOS / Linux

Development

npm install
npm run compile        # tsc → out/
npx @vscode/vsce package
code --install-extension freekit-2.4.7.vsix --force

Press F5 for an Extension Development Host.

Support

Freekit is free — no API keys, no subscriptions. If it saves you time, you can buy me a coffee:

License

Proprietary. See LICENSE.txt.

Made by builderweb · Buy me a coffee