---
id: assistant-stream
version: "0.0.36"
license: unclear
license_treatment: permissive
maintenance: active
---

# assistant-stream

License: permissive · Maintenance: active · Downloads: 91.3K/mo

## What it is and what it does

assistant-stream is a Python package that adds server-side streaming capabilities to Starlette-based backends, designed to work with the assistant-ui React library for building AI chat applications. It handles the server-side mechanics of streaming responses from language models or AI backends to connected clients in real time, complementing the React frontend's ability to display streaming messages as they arrive.

The package sits in the backend layer of an AI chat stack: you use it to stream responses from your LLM or AI service through a Starlette web server to the browser. It abstracts away the complexity of managing persistent connections and chunked response delivery, letting you focus on the business logic of generating chat responses. With only Starlette as a runtime dependency, it has minimal install friction and integrates into existing Python web applications.

Use it for:

- Stream LLM responses in real time from a Python backend to a React chat interface without waiting for the full response to generate.
- Build a production chat service where users see tokens or message chunks arrive progressively, improving perceived responsiveness.
- Integrate with Vercel AI SDK or LangGraph backends to stream tool calls and structured outputs to the frontend as they execute.
- Add streaming to an existing Starlette application without replacing the entire web framework or rewriting request handling.

## Worth the install?

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

Provides server-side streaming support for AI chat applications built with Starlette, enabling real-time message delivery to React frontends.

Yes, if you are building a Python backend for an AI chat application using Starlette and need server-side streaming. The package has low install friction, active maintenance, no known vulnerabilities, and a permissive MIT license. It is a focused tool for a specific use case (streaming responses to assistant-ui frontends) rather than a general-purpose library, so install it only if you are actually building that stack.

## Install

pip install assistant-stream
uv add assistant-stream
poetry add assistant-stream

## Installing assistant-stream

Before you install: Low friction install with a single runtime dependency (starlette). Active maintenance with recent releases; last commit 2026-08-14 and 11656 GitHub stars indicate ongoing development.

License in practice: MIT license permits commercial and private use with minimal restrictions; permissive treatment means you can integrate this into proprietary applications without licensing obligations.

Quickstart:

pip install assistant-stream

from assistant_stream import StreamingResponse
from starlette.applications import Starlette

app = Starlette()

@app.route('/chat', methods=['POST'])
async def chat(request):
    return StreamingResponse(generate_stream())

Requires Python 3.10 or later (package specifies >=3.10,<4.0).

Verify before relying:

- Whether this package provides the complete streaming infrastructure or is a thin wrapper around Starlette's native streaming capabilities.
- What specific streaming protocol or format (SSE, chunked transfer, etc.) is used for message delivery.
- Whether integration with the React assistant-ui library requires additional setup beyond this package.

## Package facts

- License: not declared (permissive)
- Python support: supports_current
- Install friction: low
- Maintenance: active
- Downloads: 91.3K/month (top 15,000 on PyPI)
- Known vulnerabilities: none known

## Tags

server-sent events streaming, ai chat backend streaming, starlette async streaming, real-time chat responses, llm response streaming python, streaming, ai-chat, starlette

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