OpenRouter
Free tierThe Unified Interface For Every Model
Free tier available·All audiences·Powered by Multi-provider (Anthropic, OpenAI, Google, and 80+ providers including ElevenLabs)·API available
Key strengths
Unified API for 500+ models across 80+ providersOpenAI-compatible APIProvider fallback and high availabilityPay-as-you-go at provider list prices with no inference markupCustom data policies and fine-grained routing controlsEdge infrastructure for low latency
Free tier + paid plans
Founded 2023
Facts verified against openrouter.ai on
Recent changes
- Plans updated: Free tier (50 req/day, 25+ free models, 4 free providers); Standard plan: pay-as-you-go with 5.5% fee on credit purchases; Business plan: pay-as-you-go with 8% fee on credit purchases; Enterprise: contract-based with invoicing. Inference billed at provider list price.
- Models updated: Multi-provider (Anthropic, OpenAI, Google, and 80+ providers including ElevenLabs)
Get alerts when OpenRouter changes
Free. We'll email you when pricing, the free tier, API access or open-source status changes.
No ratings yet
OpenRouter — Developer Documentation
Getting Started
OpenRouter's API is OpenAI-compatible, meaning any SDK or library that supports a configurable base URL works out of the box. Point your client at OpenRouter's endpoint, supply your API key, and specify the model identifier (e.g. anthropic/claude-3-5-sonnet).
Key API Concepts
- Model routing: Specify a single model or supply an ordered list of fallback models. OpenRouter handles provider selection and retry logic automatically.
- Provider controls: Per-request headers allow you to pin to specific providers, exclude providers, or enforce custom data-handling policies.
- Streaming: Server-sent events (SSE) streaming is supported, consistent with the OpenAI streaming interface.
- Unified response schema: All model responses are normalised to the OpenAI chat completion schema regardless of the underlying provider.
Integrations
- Vercel AI SDK — use OpenRouter as a custom provider within the SDK's unified interface.
- Replit — configure OpenRouter as the model backend inside Replit AI projects.
- Cline — connect Cline's agentic coding assistant to any OpenRouter-supported model.
- ElevenLabs — route to ElevenLabs voice models alongside text models through the same API.
Availability & Infrastructure
- Edge infrastructure for globally low-latency inference.
- Provider fallback ensures high availability across 80+ providers.
- 500+ models available, spanning Anthropic, OpenAI, Google, ElevenLabs, and more.
Open Source
OpenRouter's core is open source with 691 GitHub stars.
Alternatives
See all alternatives