VarwoFreeLLMAPI GitHub: Setup Routes, Keys and Free-Tier Limits
FreeLLMAPI combines the provider API keys you supply behind a local, OpenAI-compatible endpoint. Start with the official repository, choose an installation route, then check which models your own keys can actually use.
Official GitHub repository · Installation documentation · API reference
Sources checked November 2, 2026. This is a documentation-based setup guide. Varwo has not installed the router, supplied provider credentials, run inference or measured its failover performance.
What you are downloading
The project is tashfeenahmed/freellmapi, a local router and dashboard. Its README describes combining provider free tiers, routing requests and falling back when a provider is unavailable or rate limited. It lists an MIT license and positions the project for personal experimentation.
The advertised total token capacity describes a collection of provider allowances. It is not a monthly allocation granted to every installation. Your usable models and quotas depend on the accounts, keys and limits you actually have. This distinction matters more than the headline model count.
Choose one setup route
| Route | What to prepare | Where to continue |
|---|---|---|
| Desktop app | An installer matching your operating system and CPU | Follow the desktop section of the installation documentation |
| Docker Compose | Docker, Compose and a persistent data volume | Follow the documented Compose configuration and preserve the encryption key |
| Local development | Node.js 20 or later and npm | Follow the development section; its UI runs separately from the API |
The installation guide documents the server dashboard on port 3001 and the development UI on 5173. It also distinguishes the server's dashboard account from the desktop application's automatic local sign-in. Use the instructions for your chosen route rather than mixing their login steps.
After installation, add your own provider credentials on the Keys page, choose a fallback chain, then obtain the separate unified API key for your client. Preserve both the database and its encryption key when upgrading. Changing the encryption key alone does not re-encrypt stored credentials; the guide provides a separate rotation procedure.
Connect a client and check readiness
The API reference gives these connection settings:
| Setting | Documented value or source |
|---|---|
| OpenAI-compatible base URL | http://localhost:3001/v1 for the documented local server |
| Client API key | The unified key from your FreeLLMAPI dashboard |
| Automatic routing | auto follows the active fallback chain |
| Model discovery | /v1/models; available=true narrows the listing to currently usable entries |
| Local API documentation | /v1/docs |
A model appearing in the full catalog does not mean your installation can serve it. Check the reported readiness and enabled provider keys before troubleshooting the client. Once configured, try a harmless request and inspect which provider served it; that is a suggested verification step, not a test Varwo performed.
Free software, provider limits and catalog updates
As checked on November 2, the README advertises an optional $19/year live-catalog subscription. It says free installations receive the delayed monthly catalog, while premium gets changes sooner. This is a catalog-update offer, not unlimited provider inference. Confirm the current checkout terms and each provider's usage rules before spending or planning sustained workloads.
The security policy describes a single-user tool intended for localhost or a trusted network. It documents encrypted credential storage, but that is not an independent security audit. Keep the default local binding for a personal setup and protect the database, encryption key and unified client key. An internet-facing, multi-user service requires a different operational assessment.