--- title: "Deploy FreeLLMAPI | One OpenAI-Compatible API for Free LLM Tiers" description: "Route one OpenAI-style API across your free LLM keys, with fallback." category: "AI/ML" url: https://railway.com/deploy/freellmapi-server --- # Deploy FreeLLMAPI | One OpenAI-Compatible API for Free LLM Tiers Route one OpenAI-style API across your free LLM keys, with fallback. **[Deploy FreeLLMAPI | One OpenAI-Compatible API for Free LLM Tiers on Railway](https://railway.com/template/freellmapi-server)** Machine-readable deploy manifest (JSON, validated by TemplateCI): https://railway.com/deploy/freellmapi-server/manifest.json - **Creator:** Two Workspaces - **Category:** AI/ML ## Template content ### FreeLLMAPI https://raw.githubusercontent.com/two-workspaces/freellmapi/main/icon.svg - **Source:** https://github.com/two-workspaces/freellmapi - **Health check:** /api/ping - **Public domain:** Yes ## Documentation # Deploy and Host FreeLLMAPI on Railway [![Deploy on Railway](https://railway.com/button.svg)](https://railway.com/new/template/freellmapi-server?referralCode=sLjDXb&utm_medium=integration&utm_source=button&utm_campaign=freellmapi-server) FreeLLMAPI puts the free tiers of many LLM providers behind one OpenAI-compatible `/v1` endpoint. You add your own provider keys, and its router picks an available model for each request and falls over to the next one when a provider hits its limit. This template runs it on Railway with your login created in advance and a volume that keeps your data across redeploys. This is a community template, not made or supported by the FreeLLMAPI project. ## About Hosting FreeLLMAPI A new FreeLLMAPI install normally opens a first-run setup page, guarded on a public address by a one-time code from the logs. This template creates your account from `ADMIN_EMAIL` and `ADMIN_PASSWORD` before the server starts listening, so there is no setup page to claim and no code to find. Your database and encrypted keys sit on a volume at `/app/server/data`, so a redeploy or a new variable doesn't wipe them. After deploying, open the service's domain and sign in. Add provider keys on the **Keys** page, then copy the unified API key from that page's header. Point any OpenAI-compatible client at `https://‹your-domain›/v1` with that key. **Please read before relying on it.** Upstream describes the project as single-user and "for personal experimentation and learning, not production". Free tiers can change or vanish, so answers can slow down or stop. Use your own keys, and check each provider's terms; upstream publishes a [provider-by-provider review](https://github.com/tashfeenahmed/freellmapi/blob/main/docs/en/architecture/00-high-level-index.md#terms-of-service-review). This deploy has a public address, so keep the password and unified key private. ## Common Use Cases - Prototype an app against an OpenAI-style API without paying for tokens - Keep one always-on endpoint for your own editor, CLI and coding-agent tools - Try many models and providers through a single key and base URL ## Dependencies for FreeLLMAPI Hosting - Free API keys from the LLM providers you want to use ### Deployment Dependencies - [FreeLLMAPI on GitHub](https://github.com/tashfeenahmed/freellmapi) - [Install and deploy guide](https://github.com/tashfeenahmed/freellmapi/blob/main/docs/en/install/01-install.md) - [API reference](https://github.com/tashfeenahmed/freellmapi/blob/main/docs/en/api/01-rest-api.md) ### Implementation Details - **Tested from Railway:** Groq, Cerebras, Gemini, OpenRouter, Mistral, Cohere, GitHub Models, Hugging Face and others answered from a Railway service without blocking it, though that test had no real keys for most of them. Real completions came back through the router on keyless routes, including Kilo. - **Keyless routes:** a few providers work with no key but limit by IP address, and on a cloud host you may share that address with other services, so those limits can run out quickly. Your own keys avoid that. - **Keep `ENCRYPTION_KEY`:** it decrypts your stored provider keys. If it changes, you have to add them all again. - **Login variables:** `ADMIN_EMAIL` and `ADMIN_PASSWORD` apply only while no account exists. Later edits to them do nothing, and the log says so on each boot. Change your password in the dashboard, or use the one-time reset code it prints to the deploy logs. - **Same image as upstream:** the Dockerfile is one pinned `FROM` line on upstream's image, with no changes. - **Premium is optional:** upstream sells a live model catalog. Free installs get a monthly snapshot of it and need nothing else. ## Why Deploy FreeLLMAPI on Railway? Railway is a singular platform to deploy your infrastructure stack. Railway will host your infrastructure so you don't have to deal with configuration, while allowing you to vertically and horizontally scale it. By deploying FreeLLMAPI on Railway, you are one step closer to supporting a complete full-stack application with minimal burden. Host your servers, databases, AI agents, and more on Railway. ## Similar templates - [Chat Chat](https://railway.com/deploy/-WWW5r) — Chat Chat, your own unified chat and search to AI platform. - [Weaviate](https://railway.com/deploy/weaviate-1) — Vector database with hybrid search and an API key-protected API - [Möbius v0.1 Personal AI OS](https://railway.com/deploy/mobius-v01-personal-ai-os) — Self-hosted AI agent that builds and runs apps for you, with memory. Open this page in a browser: https://railway.com/deploy/freellmapi-server