Build Multi-Agent
System for |
pip install camel-ai
CAMEL-AI is an open-source community for finding the scaling laws of agents for data generation, world simulation, task automation.
We Are Finding the
Scaling Laws of Agents
Advancing Multi-Agent Research
Research with Us
We warmly invite you to use CAMEL for your impactful research.
Rigorous research takes time and resources. We are a community-driven research collective with 100+ researchers exploring the frontier research of Multi-agent Systems and intelligent agents for real-world automation. Join our ongoing projects or test new ideas with us.
CAMEL Framework Design Principles
Agents can continuously evolve via data generation and interactions with their environment, driven by reinforcement or supervised learning.
Systems with millions of agents can be built, ensuring efficient coordination, communication, and resource management at scale.
Agent context is managed as a state transition process, supporting rich, dynamic memory management over time.
Each line of code or comment serves as a prompt, ensuring that both humans and agents can interpret and extend it effectively.
Unique Capabilities of CAMEL Framework
Workforce
Model real agent workforces with roles, hierarchies, and long-horizon tasks.
CAMEL Toolkit
Batteries-included toolkit for messaging, planning, evaluation, and observability.
Connect to RL
Close the loop from interaction logs to RL and fine-tuning pipelines.
Research Ecosystem
Built on state-of-the-art CAMEL/OASIS research with open benchmarks and datasets.
Create Powerful Agentic Applications with Our Tools
Agent
Single-agent
Agent Society
Multi-agent
Data Generation
Models
Tools
Memories
Storage
Key-Value Storage:
Graph DB Storage:
Object Storage:
Vector DB Storage:
Data Loaders
Environments
Interpreters
Retrievers
Runtime
Verifier
MCP
Human in the Loop
Observe
Join Our Community
What People Are Saying
"The thing that I find really interesting with this is that it's an unbelievably good way to make synthetic data. If you're trying to create any sort of customer service or chatbot agent that communicates with the public, this allows you to make synthetic data for training and fine-tuning."
Sam Witteveen
Co-founder @ Red Dragon AI
"The CAMEL AI "Domain Expert" dataset, comprising 25,000 conversations between two GPT 3.5 Turbo agents was used as part of the training data for Teknium's OpenHermes model and the Microsoft Phi model"
Valory
Open-source framework
"Guohao Li, who designed Camel, highlights the potential of multi-agent systems to bypass traditional AI limitations, enabling tasks like phishing email generation and cyber bug development."
The Economist
Newspaper
"The essence of Camel lies in its prompt engineering, i.e., inception prompting. The prompts are actually carefully defined to assign roles, prevent flipping roles, prohibit harm and false information, and encourage consistent conversation."
Sophia Yang, Ph.D.
Head of Developer Relations @ Mistral AI
MPT-30B-Chat is a chatbot-like model for dialogue generation. It was built by finetuning MPT-30B and trained on 19.54% Camel-AI sourced data
Databricks
The Data and AI Company
"This innovative concept is set to redefine the way AI agents interact with each other and, in doing so, revolutionize the realm of conversational AI."
Yogesh Haribhau Kulkarni
AI Advisor