Pi Coding Agent

Integration overview

How it connects to OpenUI

The maintained local harness embeds @earendil-works/pi-coding-agent in a Next.js route, injects the prompt generated from OpenUI's component library, and converts Pi text, reasoning, and tool events into an OpenAI-compatible stream consumed by openAIReadableStreamAdapter(). It exposes real filesystem and shell tools, so the shipped unauthenticated harness must not be exposed to a network as-is.

The integration path

1. 1

   Describe the interface

   Generate a component prompt from the same OpenUI library that will render the response.

2. 2

   Stream structured output

   Let Pi Coding Agent own its part of the stack while OpenUI Lang travels as incremental text or mapped agent events.

3. 3

   Render and interact

   Parse the stream with the matching OpenUI runtime, render real components, and return validated actions to the application.

What stays consistent