LiteLLM

Name: LiteLLM
Author: BerriAI

by BerriAIFramework / SDKlitellm.ai

LiteLLM gives Python one call signature across providers, and a proxy that puts the same thing behind a server. qynio works with it today through the OpenAI-compatible provider: point it at https://api.qyn.io/v3/compat with your qynio key and every model in the catalogue is reachable. A native qynio/ provider is open for review upstream in BerriAI/litellm#36075.

Build with LiteLLM on qynio

LiteLLM reaches qynio over the OpenAI-compatible endpoint. Point it at https://api.qyn.io/v3/compat with your qynio key and pick a model.

Get your API key

In your code

1import os
2
3from litellm import completion
4
5response = completion(
6    model="openai/anthropic/claude-fable-5",
7    api_base="https://api.qyn.io/v3/compat",
8    api_key=os.environ["QYN_API_KEY"],
9    messages=[{"role": "user", "content": "Hello!"}],
10)

More apps on qynio

Mastra

qynio is a provider in Mastra's model router. Write a model string, set a key, and the agent runs through the gateway.

Frameworksqynio built in

Vercel AI SDK

Our provider for the Vercel AI SDK, so generateText, streamText and tool calls run on any model behind one key.

Frameworks@qynio/ai-sdk-provider

LangChain

Point LangChain's OpenAI chat model at qynio and the rest of the chain carries on unchanged.

Frameworks/v3/compat