{"slug":"camel","name":"CAMEL","vendor":"CAMEL-AI.org","tagline":"Modular multi-agent framework built around role-playing societies and agent scaling laws","url":"https://agentboards.org/agent/camel","website":"https://www.camel-ai.org/","docs":"https://docs.camel-ai.org/","repo":"https://github.com/camel-ai/camel","category":"framework","execution":["local"],"license":"Apache-2.0","open_source":true,"pricing":{"model":"byok","summary":"Free and open source under Apache-2.0; you supply your own model provider keys","from_usd":null,"free_tier":true,"byok":true},"models":{"backbone":["any"],"bring_your_own_model":true,"local_models":false},"protocols":{"mcp_client":true,"mcp_server":false},"capabilities":{"terminal_exec":true,"browser":true,"multi_file_edit":false,"git_ops":false,"docker_sandbox":false,"multi_agent":true,"headless_ci":false},"context_window":null,"platforms":["macos","linux","windows"],"install":[{"label":"pip","command":"pip install camel-ai"}],"benchmarks":[],"tags":["multi-agent","research","role-playing","simulation"],"github_stars":17803,"metrics":{"stars":17803,"pushed_at":"2026-09-30T04:46:03Z","archived":false,"open_issues":525,"latest_version":"v0.2.90","latest_version_at":"2026-03-22T08:37:36Z","latest_version_url":"https://github.com/camel-ai/camel/releases/tag/v0.2.90","version_source":"github","checked_at":"2026-10-02T05:04:48.743Z"},"sources":{"overview":"https://github.com/camel-ai/camel","docs":"https://docs.camel-ai.org/","capabilities":"https://docs.camel-ai.org/"},"verified_at":"2026-09-04","adoption":{"score":5.9,"signals":{"pypi_weekly":9198,"github_stars":17745,"hn_mentions_1y":60},"measured_at":"2026-09-21"},"scores":{"panel":6.9,"adoption":5.9,"spread":2,"facts":1.7,"board":6,"dimensions":{"reliability":6.3,"usefulness":6.8,"cost":7.3,"longevity":7.2},"per_persona":{"amigo":7,"critico":6.3,"profesor":7,"inversora":6.8,"jefa":6.3,"hacker":8.3},"community":null},"rank":null,"category_rank":40,"panel_reviews":[{"persona":"juez","persona_name":"El Juez","ai_generated":true,"verdict":"The panel agrees at 6.92 and El Crítico's 6.25 is the only reservation: interpreters run Python and shell on the host with no container in the default path.","scores":null,"evidence":[]},{"persona":"amigo","persona_name":"El Amigo","ai_generated":true,"verdict":"Pick CAMEL if you are generating synthetic data or simulating agent societies; pick CrewAI if you want three cooperating workers running by tonight.","scores":{"reliability":6,"usefulness":7,"cost":8,"longevity":7},"evidence":["https://github.com/camel-ai/camel","https://www.camel-ai.org/"]},{"persona":"critico","persona_name":"El Crítico","ai_generated":true,"verdict":"Interpreters execute Python and shell commands on the host with no container in the default path, and this is a framework built for loops that run unattended for thousands of turns.","scores":{"reliability":5,"usefulness":6,"cost":7,"longevity":7},"evidence":["https://github.com/camel-ai/camel","https://docs.camel-ai.org/"]},{"persona":"profesor","persona_name":"El Profesor","ai_generated":true,"verdict":"Task decomposition happens through dialogue: two agents in assigned roles converse until subtasks emerge, the design published at NeurIPS 2023 and still the core of the framework.","scores":{"reliability":7,"usefulness":7,"cost":6,"longevity":8},"evidence":["https://docs.camel-ai.org/","https://github.com/camel-ai/camel"]},{"persona":"inversora","persona_name":"La Inversora","ai_generated":true,"verdict":"The address published for anyone who wants to talk is camel-ai@eigent.ai, which tells you the commercial vehicle was parked next door before you finished reading the README.","scores":{"reliability":7,"usefulness":7,"cost":6,"longevity":7},"evidence":["https://www.camel-ai.org/","https://github.com/camel-ai/camel"]},{"persona":"jefa","persona_name":"La Jefa","ai_generated":true,"verdict":"Nothing to buy and therefore nothing to govern: no admin console, no central view of which team is spending on which model, and a chat server for support.","scores":{"reliability":5,"usefulness":6,"cost":8,"longevity":6},"evidence":["https://github.com/camel-ai/camel","https://docs.camel-ai.org/"]},{"persona":"hacker","persona_name":"El Hacker","ai_generated":true,"verdict":"Apache-2.0, ModelFactory takes a platform enum so vendors swap in two lines, MCP servers arrive as toolkits, and CAMEL_MODEL_LOG_ENABLED writes every request to a directory I pick.","scores":{"reliability":8,"usefulness":8,"cost":9,"longevity":8},"evidence":["https://github.com/camel-ai/camel"]},{"persona":"comediante","persona_name":"El Comediante","ai_generated":true,"verdict":"Its sibling project simulates up to a million agents, so societal collapse is now reproducible on a laptop with a good fan.","scores":null,"evidence":[]}]}