> ## Documentation Index
> Fetch the complete documentation index at: https://vibes-sdk.a7ul.com/llms.txt
> Use this file to discover all available pages before exploring further.
# Vibes Agent SDK
> TypeScript agent framework for building production-grade, type-safe AI applications and workflows, the [Pydantic AI](https://ai.pydantic.dev) way, using [Vercel AI SDK](https://sdk.vercel.ai).
## TypeScript Agent Framework, the Pydantic AI way
Build production-grade AI agents with type-safe tools, dependency injection, structured output, and first-class testing.
```ts theme={null}
import { Agent } from "@vibesjs/sdk";
import { anthropic } from "@ai-sdk/anthropic";
const agent = new Agent({
model: anthropic("claude-haiku-4-5-20251001"),
systemPrompt: "You are a helpful assistant.",
});
const result = await agent.run("What is the capital of France?");
console.log(result.output); // "Paris"
```
## Why Vibes?
1. **Type-safe tools + Dependency injection** - Every tool parameter is validated at runtime with Zod. Carry databases, HTTP clients, and config via `RunContext` through the entire call chain. No `any` types, no global state.
2. **Automatic retries + Cost control** - Vibes retries on validation failure and enforces token budgets and request limits to keep costs in check.
3. **Structured output + Streaming** - Define a Zod schema, get back a typed object or stream typed partial objects to the client as they arrive.
4. **Testing + Evals - the only way to ship AI to production** - Unit-test every agent in CI with `TestModel` and `setAllowModelRequests(false)` (no real API calls). Then go further with typed eval datasets, built-in and LLM-as-judge evaluators, and experiment runners with configurable concurrency. Evals are code - they live in your repo, run in CI, and catch regressions before they reach users.
5. **Model-agnostic** - Switch between Anthropic, OpenAI, Google, Groq, Mistral, Ollama, and 50+ providers by changing one line.
6. **OpenTelemetry observability** - Every run emits OTel spans, events, and token usage metrics. Works with Jaeger, Honeycomb, Datadog, and any OTel-compatible backend.
7. **Durable agents + MCP, AG-UI, A2A** - Run long-lived agents that survive crashes and restarts with Temporal. Connect to MCP servers and build AG-UI and A2A agents out of the box.
## Architecture
```mermaid theme={null}
flowchart TD
A[Your Code] --> B[Vibes SDK]
B --> C[AI SDK]
C --> D[Anthropic]
C --> E[OpenAI]
C --> F[Google]
C --> G[Groq / Mistral / ...]
```
## Acknowledgments
Vibes is inspired by [Pydantic AI](https://ai.pydantic.dev/) - the agent framework by Samuel Colvin and the Pydantic team. Pydantic AI showed that agent frameworks can be type-safe, testable, and dependency-injection-friendly without sacrificing simplicity. Vibes borrows its API design, agent loop architecture, and teaching philosophy, adapted for TypeScript.
Vibes uses [Vercel AI SDK](https://sdk.vercel.ai/) for LLM calls, which provides a unified interface to 50+ LLM providers. Without the AI SDK, Vibes would require maintaining its own provider integrations.
See the [Introduction](/introduction) for the full story.
Learn about Vibes' design philosophy and what inspired it
Set up Vibes in under 2 minutes
Build your first agent
