LLM
OpenAI, Anthropic and Meta in one API, with your own keys
One endpoint in the OpenAI chat completions format for the models of all three providers. Each call uses the key you registered for that provider, and switching models is changing a string.
POST /api/llm/chat/completions
Authorization: Bearer <merchant token>
{
"model": "anthropic/claude-sonnet-5-5",
"max_tokens": 1024,
"messages": [
{
"role": "system",
"content": "You summarize marketing data in English."
},
{
"role": "user",
"content": "Summarize the account's week."
}
]
}What you can do
- One LLM API for OpenAI, Anthropic and MetaCall OpenAI, Anthropic and Meta (Muse) models with the OpenAI chat completions format, using the SDK you already have.
- Switch LLM providers without changing codeSwitch between OpenAI, Anthropic and Meta by changing only the model field: same body, same response, same SDK.
- Automatic fallback between LLM modelsSend up to 3 backup models in models and the call moves to the next if the main one fails, even across different providers.
- Stream LLM responses over SSEUse stream: true and receive the model's answer piece by piece, as text/event-stream, in the OpenAI format with models from three providers.
- Function calling with OpenAI, Anthropic and MetaSend tools in the OpenAI format and get tool_calls from models that support them, with the same code for OpenAI, Anthropic and Meta.
- Send images to LLMs (vision)Send image_url parts in messages and ask the model to analyze images, in the OpenAI format and with models from three providers.
- List LLM models and pricesGET /llm/models lists the models you can call, with price per 1M tokens, limits and whether they accept tools and images.
Other features
See Connect applied to your product.
Try it for 15 days with no card, or walk our team through your use case.
