← All notes

Summary

CAMEL is a communicative-agent framework that uses role-playing and inception prompting so chat agents can cooperate on tasks with less ongoing human guidance. The authors use the setup to generate conversational data for studying multi-agent behavior, and they report studies of instruction-following cooperation. They open-sourced a library at the named GitHub repository. [S1]

Claims

• id: C1 claim: 'The authors propose a communicative agent framework named role-playing.' source: S1 evidence: "we propose a novel communicative agent framework named role-playing"
• id: C2 claim: 'They use inception prompting to guide chat agents toward task completion.' source: S1 evidence: "using inception prompting to guide chat agents toward task completion"
• id: C3 claim: 'They open-source a library at a public GitHub repository.' source: S1 evidence: "open-sourcing our library to support research on communicative agents and beyond: https://github.com/camel-ai/camel"

Verification

• Code available: Paper and comments: https://github.com/camel-ai/camel (HTTP 200 on 2026-11-11) and project site https://www.camel-ai.org (HTTP 200). Not run.
• Reproduced: not reproduced (2026-11-11). Code not run.
• Related work: Live notes 2308-08155-autogen and 2210-03629-react-reasoning-and-acting. This is the paper-verification note for CAMEL itself.
• Limits: CAMEL is an early role-playing setup and data-generation framework. This note does not treat later library features as part of the v2 paper (revised 2 Nov 2023).
• License: CC BY 4.0 on the abs page (2026-11-11; icon by-4.0.png; http://creativecommons.org/licenses/by/4.0/). Submitted 31 Mar 2023 (v1); last revised 2 Nov 2023 (v2). Sources: [S1] https://arxiv.org/abs/2303.17760, v2 (2 Nov 2023). checked_on: 2026-11-11.

When to apply

Use when looking at early role-playing multi-agent cooperation, or at datasets generated from agent-to-agent chat.

Facts

• id: F1 claim: 'CAMEL is a role-playing framework for communicative agents.' confidence: medium date: 2026-11-11 sources: [S1] license: CC BY 4.0
• id: F2 claim: 'The authors use inception prompting to guide agents toward task completion while aiming to stay consistent with human intentions.' confidence: medium date: 2026-11-11 sources: [S1] license: CC BY 4.0
• id: F3 claim: 'They released a library at https://github.com/camel-ai/camel.' confidence: medium date: 2026-11-11 sources: [S1] license: CC BY 4.0

Sources

• [S1] arXiv 2303.17760, https://arxiv.org/abs/2303.17760. License: CC BY 4.0. v2 (2 Nov 2023). Accessed 2026-11-11.

License: this note is CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). Summary by iinventedlight (AI-assisted). Paper is CC BY 4.0.

@misc{iil_note_jpxeyha3,
  author = {{'iinventedlight (AI-assisted)'}},
  title = {{CAMEL: Communicative Agents for "Mind" Exploration of Large Language Model Society}},
  howpublished = {iinventedlight Library},
  year = {2026},
  date = {2026-11-11},
  url = {https://iinventedlight.com/library/notes/2303-17760-camel/},
  note = {Permanent ID iil:note:jpxeyha3. License: CC-BY-4.0.},
}

Markdown

Spotted a mistake? Use the MCP feedback tool or email hello@iinventedlight.com.

For language models and agents: start at /llms.txt | /agents.md | /.well-known/agent-library.json. This page in Markdown: /library/notes/2303-17760-camel.md.