Compatible Providers

Many LLM providers use OpenAI-compatible /v1/chat/completions endpoints. llmock works with all of them out of the box — just point the SDK's base URL at your llmock instance.

Supported Providers

How It Works

• Most OpenAI-compatible providers send requests to /v1/chat/completions with the same JSON format — llmock already handles this natively
• Groq uses a /openai/v1/ prefix for all endpoints. llmock automatically strips the /openai prefix, so /openai/v1/chat/completions, /openai/v1/embeddings, and /openai/v1/models all work transparently
• Model names are passed through as-is — use mistral-large-latest, llama-3.3-70b-versatile, llama3.2, or any other model name in your fixtures

Mistral Configuration

Mistral's SDK uses the standard OpenAI-compatible endpoint. Point MISTRAL_API_ENDPOINT at llmock:

export MISTRAL_API_ENDPOINT="http://localhost:5555/v1"
export MISTRAL_API_KEY="mock-key"

import { Mistral } from "@mistralai/mistralai";

const client = new Mistral({
  apiKey: "mock-key",
  serverURL: "http://localhost:5555/v1",
});

Groq Configuration

Groq's SDK sends requests to /openai/v1/chat/completions (note the /openai prefix). llmock handles this automatically.

export GROQ_BASE_URL="http://localhost:5555/openai/v1"
export GROQ_API_KEY="mock-key"

import Groq from "groq-sdk";

const client = new Groq({
  apiKey: "mock-key",
  baseURL: "http://localhost:5555/openai/v1",
});

Ollama Configuration

Ollama exposes an OpenAI-compatible endpoint locally. Point the OpenAI SDK at llmock instead:

export OPENAI_BASE_URL="http://localhost:5555/v1"
export OPENAI_API_KEY="mock-key"

import OpenAI from "openai";

// Same SDK you'd use with Ollama, just different base URL
const client = new OpenAI({
  apiKey: "mock-key",
  baseURL: "http://localhost:5555/v1",
});

Together AI Configuration

export TOGETHER_BASE_URL="http://localhost:5555/v1"
export TOGETHER_API_KEY="mock-key"

vLLM Configuration

# vLLM uses the OpenAI SDK — just change the base URL
export OPENAI_BASE_URL="http://localhost:5555/v1"
export OPENAI_API_KEY="mock-key"

Example Fixture

The same fixture works for all compatible providers. Model names are passed through — match on whatever model name your code sends:

{
  "fixtures": [
    {
      "match": {
        "model": "mistral-large-latest",
        "userMessage": "hello"
      },
      "response": {
        "content": "Bonjour! How can I help?"
      }
    },
    {
      "match": {
        "model": "llama-3.3-70b-versatile",
        "userMessage": "hello"
      },
      "response": {
        "content": "Hey there! What can I do for you?"
      }
    },
    {
      "match": { "userMessage": "hello" },
      "response": {
        "content": "Hi! I'm a catch-all response."
      }
    }
  ]
}