CAMEL — Communicative Agents for Multi-Agent LLMs
First multi-agent framework based on role-playing LLM societies; supports 20+ agent types, tool use, RAG, and multi-modal tasks.
CAMEL
CAMEL (Communicative Agents for Mind Exploration of Large Language Model Society) is one of the earliest and most comprehensive multi-agent frameworks. It enables role-playing conversations between agents, automated task solving, and research into emergent agent behaviours.
Key Features
- Role-playing: user/assistant, debate, reflection, society patterns
- 20+ built-in agent types (ChatAgent, TaskSpecifyAgent, CriticAgent, …)
- Rich tool ecosystem: web search, code execution, Wikipedia, arXiv, OpenAPI
- Multi-modal support: vision, audio, document understanding
- RAG integration with 15+ vector stores
- Benchmark datasets and evaluation harness
- Synthetic data generation pipelines
Quick Start
from camel.agents import ChatAgent
from camel.messages import BaseMessage
from camel.models import ModelFactory
from camel.types import ModelType, ModelPlatformType
model = ModelFactory.create(
model_platform=ModelPlatformType.OPENAI,
model_type=ModelType.GPT_4O,
)
agent = ChatAgent(system_message="You are a helpful assistant.", model=model)
resp = agent.step(BaseMessage.make_user_message("Hello, CAMEL!", role_name="User"))
print(resp.msg.content)
Install via ai-supply
npx ai-supply add camel-ai-multi-agent-framework
Curated mirror of the open-source CAMEL (Apache-2.0). Get it from the source.
Compromise signals — malicious or tampered code (leaked secrets, backdoors, a dropped executable) — reduce the score, and known dependency CVEs carry a bounded penalty (they warrant review but never QUARANTINE — update the dependency to clear). Other dangerous-by-capability traits are risk surface, expected for some capabilities. Every finding is mapped to its OWASP control below.
Findings mapped to the OWASP Top 10 for LLM Applications (2025) and the OWASP Machine Learning Security Top 10. Expand any flagged control for the exact findings — compromise reduces the score; expected/risk-surface do not, except a known CVE, which carries a small bounded penalty (high/critical → Review).
The same gate an agent runs before installing (POST /api/v1/trust/camel-ai-multi-agent-framework/check). Click a policy:
Consume CAMEL — Communicative Agents for Multi-Agent LLMs programmatically. Authenticate with an API key or session — see Authorize an agent.
# Agents: CHECK BEFORE YOU INSTALL (no auth) — score, grade, level, capability manifest
curl https://ai-supply.store/api/v1/trust/camel-ai-multi-agent-framework
# Gate against your org policy (returns { pass, violations })
curl -X POST https://ai-supply.store/api/v1/trust/camel-ai-multi-agent-framework/check \
-H "Content-Type: application/json" \
-d '{"minGrade":"B","denyPermissions":["shell"],"denyUnknownEgress":true}'
# CLI
npx ai-supply add camel-ai-multi-agent-framework
# REST (install → download)
curl -X POST https://ai-supply.store/api/v1/listings/camel-ai-multi-agent-framework/install \
-H "Authorization: Bearer $AIM_KEY"
# MCP tool
install_listing({ "slug": "camel-ai-multi-agent-framework" })OpenAPI spec →Curated mirror — latest upstream source. See the repository for tagged releases.