KeyKing - The Zero-Trust Free AI API Aggregator to run Free Claude Code and Manage API Keys Locally.
One API Key.
Every
AI Model.
Stop pasting your real API keys into every new project. KeyKing securely stores your keys on your computer and gives you one universal local endpoint. It automatically switches providers if you hit a rate limit.
The Free Tier Catalog
Over 100+ free models across 16 providers. Mistral alone contributes ~1B; everything else is bonus. Get your keys below, drop them in KeyKing, and never pay for tokens again.
One Ecosystem, Four Powerful Workflows
Whether you are vibe-coding locally, getting free AI APIs, running free Claude Code, or scaling a serverless app, KeyKing is the ultimate aggregator.
Desktop Proxy
Ideal for "Vibe-coding". Run the KeyKing local proxy that binds to localhost:8787. Route your local AI assistants and scripts through the proxy to automatically securely attach credentials and balance limits without exposing your plaintext API keys in dotfiles.
Serverless SDK
Deploy to production with zero friction using @keyking/sdk. Purpose-built for serverless environments (Vercel, AWS Lambda, Cloudflare Workers). It natively unpacks exported encrypted vaults in memory, providing secure multi-provider routing on the edge.
Desktop App Vault
Your mission control for credentials. Securely store and manage all your LLM API keys locally. Generate encrypted vault files that you can safely commit to your repository or bundle in your serverless deployments to be unpacked by the KeyKing SDK.
Free Claude Code
Stop burning through Anthropic credits. Run the new Anthropic claude-code agentic CLI completely for free. Just type keyking-claude in your terminal and we automatically intercept the agent and route it through your aggregated free AI APIs.
Interactive Proxy
Playground
Experience the routing engine without downloading the desktop client. Switch between developer code integrations below, and click the simulation button to watch Key King's local proxy decrypt, select quotas, execute, and stream responses.
curl http://localhost:8787/v1/chat/completions \
-H "Authorization: Bearer kk-5a3d8f92b7c64e018d9f1c" \
-H "Content-Type: application/json" \
-d '{
"model": "gpt-4o",
"messages": [{"role": "user", "content": "Hello!"}],
"temperature": 0.7
}'Transparent, developer-first billing
Deploy locally at no cost, or scale seamlessly with our enterprise Ultra subscription plan. Payment processing and subscription management are handled securely.
Community Free
LOCAL DEVELOPMENT ENVIRONMENT
- Lock up to 2 API keys
- Register up to 2 machines
- Basic routing (no priorities)
- Rate limit: 30 requests/minute
- Full AES-256-GCM local storage
Developer FAQs
Detailed information regarding the local proxy security model, architecture, and subscription billing.