{"slug":"kode-agent-sdk","name":"KODE SDK","vendor":"shareAI-lab","tagline":"Event-driven TypeScript SDK for long-running agents: staged checkpoints, resume and fork, agent pools and cloud sandbox execution","url":"https://agentboards.org/agent/kode-agent-sdk","website":"https://github.com/shareAI-lab/kode-agent-sdk","docs":"https://github.com/shareAI-lab/kode-agent-sdk/blob/main/docs/en/reference/api.md","repo":"https://github.com/shareAI-lab/kode-agent-sdk","category":"framework","execution":["local","sandbox"],"license":"MIT","open_source":true,"pricing":{"model":"byok","summary":"Free and open source under MIT; you pay your model provider and, if you use one, your sandbox provider","from_usd":0,"free_tier":true,"byok":true},"models":{"backbone":["Anthropic","OpenAI","Gemini"],"bring_your_own_model":true,"local_models":false},"protocols":{"mcp_client":true,"mcp_server":false,"openapi":false},"capabilities":{"terminal_exec":true,"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":"npm","command":"npm install @shareai-lab/kode-sdk"}],"benchmarks":[],"tags":["open-source","typescript","sdk","checkpoints","sandbox","mcp"],"github_stars":401,"metrics":{"stars":401,"pushed_at":"2026-07-10T23:20:01Z","archived":false,"open_issues":5,"latest_version":null,"latest_version_at":null,"latest_version_url":null,"version_source":null,"checked_at":"2026-10-02T05:00:59.116Z"},"sources":{"overview":"https://github.com/shareAI-lab/kode-agent-sdk","install":"https://www.npmjs.com/package/@shareai-lab/kode-sdk","capabilities":"https://github.com/shareAI-lab/kode-agent-sdk","protocols":"https://github.com/shareAI-lab/kode-agent-sdk","models":"https://github.com/shareAI-lab/kode-agent-sdk","license":"https://github.com/shareAI-lab/kode-agent-sdk"},"verified_at":"2026-09-04","adoption":{"score":0.4,"signals":{"github_stars":401},"measured_at":"2026-09-28"},"scores":{"panel":6.5,"adoption":0.4,"spread":2,"facts":8.3,"board":5.9,"dimensions":{"reliability":6.7,"usefulness":6.2,"cost":7,"longevity":6.2},"per_persona":{"amigo":6.5,"critico":6,"profesor":7.3,"inversora":6.3,"jefa":5.5,"hacker":7.5},"community":null},"rank":null,"category_rank":49,"panel_reviews":[{"persona":"juez","persona_name":"El Juez","ai_generated":true,"verdict":"El Profesor and El Crítico agree the checkpoint design is the strongest thing here and split on what a resumed run is actually restoring.","scores":null,"evidence":[]},{"persona":"amigo","persona_name":"El Amigo","ai_generated":true,"verdict":"Pick KODE SDK if you are building an agent product that has to survive a crash; pick a smaller library if your runs finish inside a single request.","scores":{"reliability":7,"usefulness":6,"cost":7,"longevity":6},"evidence":["https://github.com/shareAI-lab/kode-agent-sdk","https://github.com/shareAI-lab/kode-agent-sdk/blob/main/docs/en/reference/api.md"]},{"persona":"critico","persona_name":"El Crítico","ai_generated":true,"verdict":"A checkpoint restores the agent, not the world. Files written and commands executed before the fork point stay done, so resuming replays side effects the snapshot never captured.","scores":{"reliability":6,"usefulness":6,"cost":6,"longevity":6},"evidence":["https://github.com/shareAI-lab/kode-agent-sdk","https://github.com/shareAI-lab/kode-agent-sdk/blob/main/docs/en/reference/api.md"]},{"persona":"profesor","persona_name":"El Profesor","ai_generated":true,"verdict":"A run is checkpointed across seven declared stages with one designated safe fork point, which turns resumption from a best-effort retry into a specified operation.","scores":{"reliability":8,"usefulness":7,"cost":7,"longevity":7},"evidence":["https://github.com/shareAI-lab/kode-agent-sdk/blob/main/docs/en/reference/api.md","https://github.com/shareAI-lab/kode-agent-sdk"]},{"persona":"inversora","persona_name":"La Inversora","ai_generated":true,"verdict":"398 stars, a permissive licence and a research-lab author: this is published infrastructure, which builds standing rather than revenue, and standing is not a business model.","scores":{"reliability":6,"usefulness":6,"cost":8,"longevity":5},"evidence":["https://github.com/shareAI-lab/kode-agent-sdk","https://www.npmjs.com/package/@shareai-lab/kode-sdk"]},{"persona":"jefa","persona_name":"La Jefa","ai_generated":true,"verdict":"Nothing to license across sixty engineers, but execution happens in a third-party remote sandbox, which is a new vendor, a new questionnaire and a new place my code goes.","scores":{"reliability":5,"usefulness":5,"cost":6,"longevity":6},"evidence":["https://github.com/shareAI-lab/kode-agent-sdk","https://www.npmjs.com/package/@shareai-lab/kode-sdk"]},{"persona":"hacker","persona_name":"El Hacker","ai_generated":true,"verdict":"MIT, one npm install, and the OpenAI-compatible provider takes a custom base URL, so DeepSeek, GLM, Qwen, Minimax or a router all work without waiting for support.","scores":{"reliability":8,"usefulness":7,"cost":8,"longevity":7},"evidence":["https://github.com/shareAI-lab/kode-agent-sdk","https://www.npmjs.com/package/@shareai-lab/kode-sdk"]},{"persona":"comediante","persona_name":"El Comediante","ai_generated":true,"verdict":"Agents collaborate through an AgentPool and Room messaging, the first group chat where nobody is quietly reading something else.","scores":null,"evidence":[]}]}