Thin LLM streaming for chirp.
Provider string in, typed async iterables out. Not a framework.
Basic usage::
from chirp.ai import LLM, AgentRun
llm = LLM("anthropic:claude-sonnet-4-20250514")
# Generate text
text = await llm.generate("Explain quantum computing simply")
# Stream unified events
async for event in llm.stream_events("Analyze this document:"):
...
# Agent loop with tools
run = AgentRun(llm, app.tools)
async for event in run.stream("What's the weather?"):
...
Requireshttpx::
pip install chirp[ai]
ai
| Name | Type | Default | Description |
|---|---|---|---|
type
|
|
— | |
qualified_name
|
|
— | |
element_type
|
|
— | |
description
|
|
— | |
source_file
|
|
— | |
line_number
|
|
— | |
is_autodoc
|
|
— | |
autodoc_element
|
|
— | |
_autodoc_template
|
|
— | |
_autodoc_url_path
|
|
— | |
_autodoc_page_type
|
|
— | |
title
|
|
— | |
doc_content_hash
|
|
— |
View source · /home/runner/work/chirp/chirp/site/../src/chirp/ai/__init__.py:1