provider:model in your config.
Quick Start
- Set the API key or login via OAuth:
- Reference the model in your agent config:
Cloud Providers
- Anthropic (Claude) — industry-leading coding and reasoning
- OpenAI (GPT, o-series) — multimodal flagship and reasoning models
- Google Gemini — massive context windows (1M+ tokens)
- Google Vertex AI — enterprise Gemini with VPC-SC, CMEK
- Amazon Bedrock — Claude, Llama, Mistral through AWS
- Azure OpenAI — OpenAI models with Azure compliance
- Groq — ultra-fast inference on custom LPU hardware
- Cerebras — wafer-scale inference, extremely fast
- Mistral — code-specialized Codestral and Devstral
- xAI (Grok) — Grok models, competitively priced
Gateways & Aggregators
- OpenRouter — meta-gateway to hundreds of models (DeepSeek, Cohere, etc.)
- Vercel AI Gateway — unified routing with observability
- LiteLLM — proxy that unifies 100+ providers behind one endpoint
Subscription & OAuth
- GitHub Copilot — use your Copilot subscription
- OpenCode — free provider, no API key needed
Regional & Specialty
- Hugging Face — open models via HF Inference API
- MiniMax — global and China endpoints
- Kimi / Moonshot — code-specialized K2 model
- ZAI / GLM — Zhipu AI, strong bilingual Chinese/English
Local & Self-Hosted
Run models on your own hardware — zero API cost, full data privacy.- Ollama — easiest way to run local models
- vLLM — high-throughput production inference
- LM Studio — desktop GUI with built-in server
- TGI — Hugging Face’s production inference server
openai-completions API mode and support auto-discovery via polpo models scan.
For custom provider setup details, API compatibility modes, and model definitions, see Custom Providers.
Quick Reference
| Provider | ID | Env Variable | Type |
|---|---|---|---|
| Anthropic | anthropic | ANTHROPIC_API_KEY | Cloud |
| OpenAI | openai | OPENAI_API_KEY | Cloud |
| Google Gemini | google | GEMINI_API_KEY | Cloud |
| Google Vertex | google-vertex | GOOGLE_CLOUD_PROJECT + ADC | Cloud |
| Amazon Bedrock | amazon-bedrock | AWS_ACCESS_KEY_ID | Cloud |
| Azure OpenAI | azure-openai-responses | AZURE_OPENAI_API_KEY | Cloud |
| Groq | groq | GROQ_API_KEY | Cloud |
| Cerebras | cerebras | CEREBRAS_API_KEY | Cloud |
| Mistral | mistral | MISTRAL_API_KEY | Cloud |
| xAI | xai | XAI_API_KEY | Cloud |
| OpenRouter | openrouter | OPENROUTER_API_KEY | Gateway |
| Vercel AI Gateway | vercel-ai-gateway | AI_GATEWAY_API_KEY | Gateway |
| GitHub Copilot | github-copilot | COPILOT_GITHUB_TOKEN | OAuth |
| Hugging Face | huggingface | HF_TOKEN | Cloud |
| MiniMax | minimax | MINIMAX_API_KEY | Cloud |
| MiniMax (CN) | minimax-cn | MINIMAX_CN_API_KEY | Cloud |
| Kimi | kimi-coding | KIMI_API_KEY | Cloud |
| ZAI | zai | ZAI_API_KEY | Cloud |
| OpenCode | opencode | OPENCODE_API_KEY (optional) | Free |
| Ollama | custom | Not required | Local |
| vLLM | custom | Not required | Local |
| LM Studio | custom | Not required | Local |
| LiteLLM | custom | LiteLLM master key | Gateway |
| TGI | custom | Not required | Local |