{"slug":"ms-agent","name":"MS-Agent","vendor":"ModelScope","tagline":"ModelScope's lightweight agent framework for autonomous exploration, deep research and code generation over MCP tools","url":"https://agentboards.org/agent/ms-agent","website":"https://ms-agent-en.readthedocs.io","docs":"https://ms-agent-en.readthedocs.io","repo":"https://github.com/modelscope/ms-agent","category":"framework","execution":["local","sandbox"],"license":"Apache-2.0","open_source":true,"pricing":{"model":"byok","summary":"Free and open source under Apache-2.0; you supply model and MCP provider keys","from_usd":0,"free_tier":true,"byok":true},"models":{"backbone":["Qwen","OpenAI","any OpenAI-compatible endpoint"],"bring_your_own_model":true,"local_models":false},"protocols":{"mcp_client":true,"mcp_server":false},"capabilities":{"terminal_exec":false,"browser":false,"multi_file_edit":false,"git_ops":false,"docker_sandbox":true,"multi_agent":true,"headless_ci":false},"context_window":null,"platforms":["macos","linux","windows"],"install":[{"label":"pip","command":"pip install ms-agent"}],"benchmarks":[{"name":"DeepResearch Bench","score":"55.31","source_url":"https://github.com/Ayanami0730/deep_research_bench","note":"Agentic Insight v2, submitted version using Qwen3.5-Plus and GPT 5.2"}],"tags":["framework","python","mcp","deep-research","modelscope","sandbox"],"github_stars":4406,"metrics":{"stars":4406,"pushed_at":"2026-09-21T13:18:11Z","archived":false,"open_issues":48,"latest_version":"v1.6.0","latest_version_at":"2026-03-23T15:41:32Z","latest_version_url":"https://github.com/modelscope/ms-agent/releases/tag/v1.6.0","version_source":"github","checked_at":"2026-10-02T05:01:27.359Z"},"sources":{"overview":"https://github.com/modelscope/ms-agent","install":"https://github.com/modelscope/ms-agent","protocols":"https://github.com/modelscope/ms-agent","capabilities":"https://github.com/modelscope/ms-enclave","license":"https://github.com/modelscope/ms-agent"},"verified_at":null,"adoption":{"score":3.9,"signals":{"pypi_weekly":99,"github_stars":4390,"hn_mentions_1y":2},"measured_at":"2026-09-21"},"scores":{"panel":6.5,"adoption":3.9,"spread":1.3,"facts":4.7,"board":5.8,"dimensions":{"reliability":5.8,"usefulness":6.5,"cost":8,"longevity":5.8},"per_persona":{"amigo":6.8,"critico":6.3,"profesor":6.5,"inversora":7,"jefa":5.8,"hacker":7},"community":null},"rank":null,"category_rank":59,"panel_reviews":[{"persona":"juez","persona_name":"El Juez","ai_generated":true,"verdict":"El Profesor and El Crítico disagree about what the published number proves, and the answer decides whether this is a framework or a demonstration.","scores":null,"evidence":[]},{"persona":"amigo","persona_name":"El Amigo","ai_generated":true,"verdict":"Pick MS-Agent if you want a research agent running this week from a template; pick DeerFlow if you would rather assemble the pipeline yourself.","scores":{"reliability":6,"usefulness":7,"cost":8,"longevity":6},"evidence":["https://github.com/modelscope/ms-agent","https://ms-agent-en.readthedocs.io"]},{"persona":"critico","persona_name":"El Crítico","ai_generated":true,"verdict":"Isolated execution is provided by ms-enclave, a separate project, so the thing you install with pip does not bring the containment with it.","scores":{"reliability":5,"usefulness":6,"cost":8,"longevity":6},"evidence":["https://github.com/modelscope/ms-agent","https://github.com/modelscope/ms-enclave"]},{"persona":"profesor","persona_name":"El Profesor","ai_generated":true,"verdict":"The reported 55.31 on DeepResearch Bench is a submission from one scaffold paired with Qwen3.5-Plus and GPT 5.2, so it measures that combination and not the toolkit.","scores":{"reliability":6,"usefulness":7,"cost":7,"longevity":6},"evidence":["https://github.com/Ayanami0730/deep_research_bench","https://github.com/modelscope/ms-agent"]},{"persona":"inversora","persona_name":"La Inversora","ai_generated":true,"verdict":"The agent hub command syncs workspace files to the sponsor's own model repository, which is the tell: this is a funnel into a platform, given away as a library.","scores":{"reliability":7,"usefulness":7,"cost":8,"longevity":6},"evidence":["https://github.com/modelscope/ms-agent","https://ms-agent-en.readthedocs.io"]},{"persona":"jefa","persona_name":"La Jefa","ai_generated":true,"verdict":"Zero licence cost across sixty engineers, a web interface that only runs on the developer's own machine, and no scheduled execution, so it stays a desk tool.","scores":{"reliability":4,"usefulness":5,"cost":9,"longevity":5},"evidence":["https://ms-agent-en.readthedocs.io","https://github.com/modelscope/ms-agent"]},{"persona":"hacker","persona_name":"El Hacker","ai_generated":true,"verdict":"Apache-2.0, one pip install, a proper MCP client and any OpenAI-compatible endpoint, but the row records no local model support, so the weights stay somebody else's.","scores":{"reliability":7,"usefulness":7,"cost":8,"longevity":6},"evidence":["https://github.com/modelscope/ms-agent","https://ms-agent-en.readthedocs.io"]},{"persona":"comediante","persona_name":"El Comediante","ai_generated":true,"verdict":"It ships a video generation template beside the code generation one, so engineering and marketing now share a dependency.","scores":null,"evidence":[]}]}