makecindy
Featured
cindy
Cindy is an open-source AI agent that runs locally on your machine, integrates multiple AI harnesses and models, and provides memory, skills, and automation to perform real work in your projects and apps.
AI repository intelligence
FR-AI / ANALYSISWhy it is worth attention
It solves the pain of managing multiple free-tier AI API keys by providing a self-contained, zero-infrastructure gateway with real-time health probing, intelligent routing, and a Provider Radar that discovers new free-tier offers—all without sending your keys anywhere or requiring a vendor account.
Who it is for
- Developers juggling multiple free-tier AI API keys
- AI agent and tool users who want reliable fallback across providers
- Privacy-conscious users who want to keep keys local and avoid hosted gateways
- Hobbyists and researchers exploring many AI providers and looking to benchmark them from their own connection
Use cases
- Centralize all your OpenAI-compatible API keys behind a single local endpoint for any client (CLI, SDK, editor plugin).
- Automatically fail over to the healthiest key when one rate-limits, times out, or returns an auth error.
- Discover new free-tier or trial inference providers via the built-in Provider Radar.
- Benchmark provider performance (TTFT, throughput, jitter) from your own machine without waiting for scheduled probes.
Strengths
- Zero infrastructure: no Docker, no Postgres, no YAML—just install and paste keys.
- Keys never leave your machine: encrypted with AES-256-GCM under your OS keychain; no telemetry, no account, no hosted control plane.
- Active health probing with per-key metrics (TTFT, latency, throughput, jitter) and a 9-strategy router with a live preview of candidate order.
- Provider Radar automatically discovers new free-tier offers from GitHub, Hacker News, and pinned feeds, with one-click addition.
Considerations
- Prebuilt installers are not yet published; you need a Rust toolchain and Node.js to build from source.
- Non-OpenAI adapters (Anthropic, Gemini) only translate text; tool-call and vision passthrough are not yet implemented.
- The gateway binds to loopback and is designed for a single user on a personal machine—not suitable for multi-user or production environments.
README quick start
Quick start
Description
Desktop control room for your AI API keys: encrypted key store, background health probes, 9 routing strategies, and a local OpenAI-compatible failover proxy. Tauri v2 + Vue 3.
Related repositories
Similar projects matched by category, topics, and programming language.
deerwork-ai
Featured
deer-workflow
An open-source Dynamic Workflow runtime that combines deterministic TypeScript orchestration with replaceable Agent runtimes.
gavamedia
Featured
deltafin
Deltafin is a research project that runs the 2.8-trillion-parameter Mixture-of-Experts model Kimi K3 on a single Apple Silicon Mac (e.g., M1 Max with 64 GB) at about 16 seconds per token, using exact, reproducible inference with local or streaming expert loading.