⭐️ Love Smoke Monkey Harness? Support our open source project! Star on GitHub ⭐ • 🌐 Interactive Agent Simulator: smoke-monkey-harness.vercel.app ↗
🐒 Smoke Monkey Harness v0.1.3
Quickstart API Reference npm GitHub
Docs Hub / Reference / 13. 18 Supported LLM Matrix

Providers

⭐️ Love Smoke Monkey Harness? Please support the project with a star on GitHub! 🌐 Interactive Simulator: Test the agent loop and stream live at smoke-monkey-harness.vercel.app

Providers are small OpenAI-compatible adapters. Set provider and pass the key via apiKey (string) or a resolver (provider, userId?) => key so secrets never live in code. Constructor is sk-... / nvapi-... style keys held in env vars.

provider env var base URL / notes
openai OPENAI_API_KEY https://api.openai.com/v1
openrouter OPENROUTER_API_KEY https://openrouter.ai/api/v1
nvidia (default) NVIDIA_API_KEY https://integrate.api.nvidia.com
xai XAI_API_KEY https://api.x.ai/v1
gemini GEMINI_API_KEY https://generativelanguage.googleapis.com
opencode OPENCODE_API_KEY OpenAI-compatible via the opencode gateway
omniroute OMNIROUTE_API_KEY https://api.getomni.app/openai/v1
huggingface HUGGINGFACE_API_KEY https://router.huggingface.co/v1 (Inference Providers router)
deepseek DEEPSEEK_API_KEY https://api.deepseek.com/v1
qwen QWEN_API_KEY (or DASHSCOPE_API_KEY) https://dashscope.aliyuncs.com/compatible-mode/v1
zai ZAI_API_KEY https://api.z.ai/api/paas/v4 (GLM)
moonshot MOONSHOT_API_KEY https://api.moonshot.ai/v1 (Kimi)
mistral MISTRAL_API_KEY https://api.mistral.ai/v1
cohere COHERE_API_KEY https://api.cohere.com/v1
groq GROQ_API_KEY https://api.groq.com/openai/v1
together TOGETHER_API_KEY https://api.together.xyz/v1
fireworks FIREWORKS_API_KEY https://api.fireworks.ai/inference/v1
cerebras CEREBRAS_API_KEY https://api.cerebras.ai/v1
ollama — http://localhost:11434 (local, no key)

Any OpenAI-compatible endpoint can be overridden with options.baseUrl.

Reference provider: NVIDIA

nvidia/nemotron-3-super-120b-a12b is the tested model (tool calls + streaming; reasoning tokens surface as reasoning_content):

ts 
const agent = createAgent({
  provider: 'nvidia',
  model: 'nvidia/nemotron-3-super-120b-a12b',
  workspacePath: process.cwd(),
});

Stream on text.delta:

ts 
agent.on('text.delta', (e) => process.stdout.write(e.data.delta));

Ignoring key resolution

Omit apiKey for local ollama. If a provider is called with no key, the loop fails with a clear message telling you which env var is missing.

Resolution order for apiKey

  1. apiKey function — resolve per call with (provider, userId).
  2. apiKey string.
  3. Provider's env var (NVIDIA_API_KEY, etc.).

userId flows into key resolution and permission requests, so a single server process can serve many users.