Skip to main content

Distri

Distri lets you embed an AI harness that can drive product experiences. Most AI agents stop at the chat box. They talk. A Distri agent reaches into your product, calls your functions, and does the work where your users already are: it fills the form, reconciles the ledger, edits the record, and streams every step back to your UI.

You describe the agent in a markdown file, hand it a few of your own functions, and run it from the terminal, as an API, or embedded in your app.

Want to see it first? The samples are recorded runs of real agents driving real product UIs.

Three steps

1. Define

An agent is a markdown file: instructions, model, and the tools it can use. No framework, no boilerplate.

2. Connect

Expose your product's functions as tools. The agent decides when to call them. Your code decides what they do.

3. Embed

Render a chat with @distri/react or @distri/angular and the agent drives your UI, streaming as it goes.

---
name = "form_filler"
description = "Fills out forms from natural conversation"
max_iterations = 5

[tools]
external = ["*"]

[model_settings]
model = "gpt-4.1-mini"
---

# ROLE
You fill out forms based on what the user describes.
Extract the fields and fill them accurately.

That file is the whole agent. Hand it your app's functions and it fills the form. Getting Started takes you from install to a running agent in a couple of minutes.

Where to find things

Features

  • Built in Rust, fast and predictable.
  • A2A JSON-RPC compatible.
  • Use MCP servers.
  • Implement client-side tools that can drive your product.
  • Integrate the agent harness into your app through APIs or client SDKs.
  • All client libraries are MIT licensed. The server is fair-code licensed.