UnisonAI Multi-Agent Framework provides a flexible, light-weight experience and extensible environment for creating and coordinating multiple AI agents.
Project description
UnisonAI
UnisonAI is a flexible Python framework to build and manage multiple AI agents—either solo or in clans. Each agent is powered by the LLM of your choice, with agent-to-agent (A2A) communication capabilities.
✨ Key Features
-
🔗 Multi-LLM Support: OpenAI, Grok, Gemini, Cohere, Anthropic, and more
-
🧩 Modular & Extensible: Add your own tools, logic, and models
-
🤖 Single or Multi-Agent: Solo agents or collaborative Clan agents
-
⛨️ Error Handling: Built-in retries and format repairs
-
📃 Clear Docs & Examples: Easy to start and extend
-
⚡ Production Ready: Ideal for real-world automation & chatbots
🚀 Installation
pip install unisonai
Requires Python >=3.10 and <3.13
🧠 Core Concepts
Single_Agent
Run focused tasks with isolated agents using any LLM and tools.
Agent (for Clans)
A specialized agent that communicates and collaborates with others.
Clan
A group of agents working together with a manager and a shared goal.
Tool System
Add your own pluggable tools with parameters and logic.
🔧 Usage Example (Single Agent)
from unisonai import Single_Agent
from unisonai.llms import Gemini
from unisonai.tools.websearch import WebSearchTool
from unisonai import config
config.set_api_key("gemini", "your-api-key")
agent = Single_Agent(
llm=Gemini(model="gemini-2.0-flash"),
identity="Web Explorer",
description="Agent to fetch stock prices",
tools=[WebSearchTool],
history_folder="history",
output_file="output.txt"
)
agent.unleash(task="Get the current price of Apple stocks")
🗺 Usage Example (Clan Agent)
from unisonai import Agent, Clan
from unisonai.llms import Gemini
from unisonai.tools.websearch import WebSearchTool
from unisonai import config
config.set_api_key("gemini", "your-api-key")
planner = Agent(...) # Define your clan leader agent
web_agent = Agent(...)
time_agent = Agent(...)
clan = Clan(
clan_name="Travel Experts",
manager=planner,
members=[web_agent, time_agent],
shared_instruction="Plan a 7-day budget trip in India",
goal="Create an itinerary under 10,000 INR",
history_folder="trip_history",
output_file="trip_plan.txt"
)
clan.unleash()
🔑 API Key Setup
from unisonai import config
config.set_api_key("gemini", "your-gemini-key")
config.set_api_key("openai", "your-openai-key")
You can also use environment variables:
export GEMINI_API_KEY="your-key"
❓ FAQ
What is UnisonAI?
A multi-agent AI framework using your favorite LLMs.
Why Clans?
Break tasks into parts and assign agents to solve each collaboratively.
Can I use custom models/tools?
Yes! Just extend the BaseLLM or BaseTool classes.
Where are logs stored?
In the folder you define using history_folder.
What can I build?
Chatbots, planners, assistants, automation systems, and more.
👥 Contributors & License
-
Lead Developer: @E5Anant
-
Contributors: @AnonymousCoderArtist, @OEvortex
License: Apache 2.0
GitHub: https://github.com/UnisonAIInc/UnisonAI
UnisonAI — Orchestrate the Future of Multi-Agent AI
Project details
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
Filter files by name, interpreter, ABI, and platform.
If you're not sure about the file name format, learn more about wheel file names.
Copy a direct link to the current filters
File details
Details for the file unisonai-0.7.tar.gz.
File metadata
- Download URL: unisonai-0.7.tar.gz
- Upload date:
- Size: 52.1 kB
- Tags: Source
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/6.1.0 CPython/3.12.9
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 |
344212c89381c10f54e8a854d0005371559a2b3e8cd5f82cb5c607788a79d447
|
|
| MD5 |
ba66315e4c540001bea6a1a04eaef661
|
|
| BLAKE2b-256 |
3debed61a3cb5b649b0a383893a2d0487c8df43a14b11c6d221e7fcf660decff
|
File details
Details for the file unisonai-0.7-py3-none-any.whl.
File metadata
- Download URL: unisonai-0.7-py3-none-any.whl
- Upload date:
- Size: 60.4 kB
- Tags: Python 3
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/6.1.0 CPython/3.12.9
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 |
0ce851e64219e673995667ecdab0f543373f17ebffa9277f5e65c04422edb53a
|
|
| MD5 |
8004c32e1c1d47218bcb261566d7d7d5
|
|
| BLAKE2b-256 |
7b9d99a7b3afccc2af4f737625a88e38707a0758fb593411f49393cc04e7912b
|