vercel-labs ai-python
COMMUNITYLABSCO SUMMARY
The package (vercel-labs/ai-python) installs with uv add ai, and OpenAI- or Anthropic-compatible providers load lazily behind optional extras (ai[openai], ai[anthropic]) rather than being bundled by default. Its ai.get_model accepts a bare model id, which defaults to routing through Vercel's AI Gateway, or an explicit openai:/anthropic: prefix for a direct provider connection. The README's own examples cover streaming text, structured output through Pydantic models, and provider-side built-in tools.
This fits Python developers building LLM applications who want one library across models and providers rather than a provider-specific SDK. It needs a working API key or Gateway credential before anything in it runs; there is nothing here for someone without one.
READ THE FULL ANALYSIS
Two of the three skills we have ever tracked for this repository are already dead, leaving one. We don't have a stated reason for the other two disappearing; what remains is the single ai skill described above.
WHAT'S INSIDE
1 showing · 1 totalai
One way to call any AI model from Python, whether what you want back is a single answer arriving word by word or an agent that keeps calling your own functions until the job is finished.
HOW TO GET IT
npx skills add vercel-labs/ai-pythonnpx skills add vercel-labs/ai-python --skill <name> --full-depthPick the skill name from the Skills tab — each entry there installs independently.