---
id: agent-framework-devui
version: "1.0.0b260813"
license: unclear
license_treatment: permissive
maintenance: active
---

# agent-framework-devui — Debug UI for Microsoft Agent Framework with OpenAI-compatible API server.

License: permissive · Maintenance: active · Downloads: 713.6K/mo

## What it is and what it does

DevUI is a sample application interface that provides both a web UI and an OpenAI-compatible REST API for running agents and workflows. It supports directory-based discovery of agent definitions, in-memory entity registration, and includes a curated gallery of sample entities to help new users get started. The UI runs on localhost with token-based authentication and can be launched either programmatically via the serve() function or from the command line with the devui CLI tool.

The package is built on fastapi and uvicorn and integrates with opentelemetry-sdk for trace visibility. It is explicitly positioned as a development and testing tool, not a production deployment platform. It supports multi-turn conversations, OpenAI proxy mode for direct model testing, and automatic cleanup of resources like MCP tool connections.

Use it for:

- Test and debug agents locally during development without building a custom UI.
- Discover and run pre-built agent examples to learn the framework.
- Expose agents via an OpenAI-compatible API for integration testing or prototyping.
- View OpenTelemetry traces of agent execution to troubleshoot behavior.
- Run multi-turn conversations with agents while managing message history automatically.

## Worth the install?

AI-flagged interpretation of the facts above — verify before relying.

Provides a lightweight web UI and OpenAI-compatible API server for discovering, registering, and running agents and workflows built with the Microsoft Agent Framework.

Yes, if you are developing or prototyping agents and want a quick, low-friction way to test them locally. The permissive MIT license, low install friction, and active maintenance make it a practical development tool. However, do not use it for production—it is explicitly a sample app, and the documentation recommends building your own custom interface for production deployments.

## Install

pip install agent-framework-devui
uv add agent-framework-devui
poetry add agent-framework-devui

## Installing agent-framework-devui

Before you install: Low install friction with a pure-Python wheel distribution. The package is actively maintained with a recent release and depends on established libraries (fastapi, uvicorn, opentelemetry-sdk, openai, agent-framework-core).

License in practice: Licensed under MIT (permissive), allowing use, modification, and distribution with minimal restrictions.

Quickstart:

pip install agent-framework-devui --pre

from agent_framework_devui import serve
from agent_framework_core import Agent

agent = Agent(name="MyAgent")
serve(entities=[agent], auto_open=True)

Requires Python 3.10 or later. OpenAI API credentials must be configured for agent execution.

Verify before relying:

- Whether the sample entity gallery examples are comprehensive enough for typical development workflows.
- Performance characteristics when running many agents or workflows concurrently through the UI.
- Whether MCP tool connection cleanup is reliable across all shutdown scenarios.
- Production readiness and scalability beyond the stated sample app scope.

## Package facts

- License: not declared (permissive)
- Python support: supports_current
- Install friction: low
- Maintenance: active
- Downloads: 713.6K/month (top 15,000 on PyPI)
- Known vulnerabilities: none known

## Tags

agent framework debug ui, agent testing interface, workflow runner web ui, openai compatible agent api, agent framework devui, local agent development server, agent discovery and execution, agent-framework, development-tool, openai-compatible

[View on SkillFed](https://skillfed.io/packages/agent-framework-devui) · [View on PyPI](https://pypi.org/project/agent-framework-devui/)