vercel-labsvercel-labsOfficialยท1 Agent Skills Included

ai-python

Build Python LLM apps, agents, and tool loops

Builds Python applications that call language, image, video, speech, and embedding models through one consistent interface. Runs agent loops with tool calling, human approvals, subagents, and durable execution without writing boilerplate orchestration code. Removes provider lock-in by routing OpenAI, Anthropic, and gateway models through a single API. Includes testing utilities and telemetry to verify model behavior before production deployment.
npx skills add vercel-labs/ai-python --all -g -y

All Skills in This Repository (1)

Pure Emerald Level Indicators

Frequently Asked Questions

FAQPage Schema
How to install ai-python?โ–ผ

Run `npx skills add vercel-labs/ai-python --all -g -y` in your terminal to install this skill globally.

How to call LLMs from Python?โ–ผ

Use `ai.get_model()` with a provider ID and `ai.stream()` to send messages and receive streamed responses. It works with OpenAI, Anthropic, and Vercel AI Gateway through one consistent API.

How to build a tool-calling agent in Python?โ–ผ

Decorate any Python function with `@ai.tool` and pass it to `ai.Agent`, which runs the loop, executes tools, and manages message history automatically.

Does ai-python support human approval steps?โ–ผ

Yes. Hooks let an agent pause for external input such as human approval, then resume with the recorded decision, including in serverless environments.

Can I test agent behavior without real API calls?โ–ผ

Yes. The SDK includes deterministic testing utilities so you can verify model interactions, tool calls, and agent loops without hitting live providers.

Related Repositories in Software Engineering

View All in Software Engineeringโ†’