Models & Providers
Bring your own model — cloud or local, with fallback.
PoorMad is provider-agnostic. Bring your own model — cloud or local.
Cloud providers
| Provider | Set in config |
|---|---|
| OpenAI | provider: openai |
| Anthropic | provider: anthropic |
| Google Gemini | provider: gemini |
| Groq / OpenRouter / others | OpenAI-compatible base URLs |
Local models
Run a local server (Ollama, llama.cpp, vLLM) and point PoorMad at it:
# config.yaml
provider: ollama
model: llama3.1:8b
base_url: http://127.0.0.1:11434
API keys
Keys are secrets — store them in .env, never in config.yaml:
POORMAD_API_KEY=sk-...
Fallback chain
You can define a fallback order so a failed provider rolls over to the next:
provider: openai fallback: - groq - ollama