Python Bindings for Ambi

A Python native extension for the Ambi AI Agent framework.

Installation

From PyPI (recommended)

pip install ambi-python

Build from source

Requires maturin and a Rust toolchain.

# Install maturin
pip install maturin

# Build and install locally
cd bindings/python
maturin develop --release

Then import as:

from ambi import Agent, AgentState, Pipeline, LLMEngineConfig

Build & Publish

cd bindings/python

# Build wheel
maturin build --release

# Publish to PyPI
maturin publish --username __token__ --password pypi-xxxxx

The built .whl files will be in target/wheels/.

Examples

Run any example from the bindings/python directory:

cd bindings/python

# Basic cloud-based chat
OPENAI_API_KEY=sk-... python examples/chat-cloud.py

# Streaming output
OPENAI_API_KEY=sk-... python examples/chat-stream.py

# Tool calling
OPENAI_API_KEY=sk-... python examples/custom-tool.py

# Custom chat template
OPENAI_API_KEY=sk-... python examples/custom-template.py

# Custom Python LLM engine
python examples/custom-engine.py

# Memory eviction
OPENAI_API_KEY=sk-... python examples/memory-eviction.py

API Reference

Python	JS equivalent	Description
`Agent.make(config)`	`Agent.make(config)`	Create an async agent
`agent.preamble("...")`	`agent.preamble("...")`	Set system prompt
`agent.template("chatml")`	`agent.template(Chatml)`	Set template type
`agent.custom_template(...)`	`agent.customTemplate(...)`	Custom template
`agent.add_tool(...)`	`agent.tool(...)`	Register a tool
`agent.with_standard_formatting()`	`agent.withStandardFormatting()`	Enable standard formatting
`agent.with_eviction_strategy(...)`	`agent.withEvictionStrategy(...)`	Set eviction strategy
`agent.count_tokens(...)`	`agent.countTokens(...)`	Count tokens
`AgentState("id")`	`new AgentState("id")`	Create session state
`LLMEngineConfig.openai(...)`	`LLMEngineConfig.openai(...)`	OpenAI engine
`LLMEngineConfig.custom(...)`	`LLMEngineConfig.custom(...)`	Custom engine
`Pipeline.chat_runner(n)`	`Pipeline.chatRunner(n)`	Chat runner
`Pipeline.custom(...)`	`Pipeline.custom(...)`	Custom pipeline
`runner.chat(...)`	`runner.chat(...)`	Sync chat
`runner.chat_stream(...)`	`runner.chatStream(...)`	Streaming chat
`resolve_request(id, result)`	`resolveRequest(id, result)`	Resolve async callback
`resolve_pipeline_request(...)`	`resolvePipelineRequest(...)`	Resolve pipeline callback

Release files for ambi-python 0.1.1

For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.

Built distribution (wheel)

Table of built distributions (wheels) for ambi-python 0.1.1
File	Interpreter	ABI	Platform	Reset
ambi_python-0.1.1-cp312-cp312-win_amd64.whl 4.2 MB May 11, 2026	CPython 3.12	CPython 3.12	Windows x86-64	Details

Release files / ambi_python-0.1.1-cp312-cp312-win_amd64.whl

Download URL	ambi_python-0.1.1-cp312-cp312-win_amd64.whl
Size	4.2 MB
Tags	CPython 3.12 Windows x86-64
SHA-256 checksum How to use checksums	`f0508ffd0fa8f171d849c397071650c64a178f6f67c72c501e4a57ca901a21d2`
BLAKE2b-256 checksum How to use checksums	`8154506a7d9469efdc1c30346f7755ec0a8a7a193626a6e54ee80fb232e47602`
Upload date	May 11, 2026
Uploaded using Trusted Publishing? What is trusted publishing?	No
Uploaded via	`twine/6.2.0 CPython/3.12.0`