{
  "$type": "site.standard.document",
  "bskyPostRef": {
    "cid": "bafyreidxldocyfrejxwajqnlpgsv22vpzabj7fx2r3hshun3jppgjy3maq",
    "uri": "at://did:plc:pgryn3ephfd2xgft23qokfzt/app.bsky.feed.post/3mitu775egkp2"
  },
  "path": "/t/built-remoroo-an-agent-for-long-running-autoresearch-workflows/175027#post_1",
  "publishedAt": "2026-04-06T15:54:20.000Z",
  "site": "https://discuss.huggingface.co",
  "tags": [
    "http://remoroo.com/"
  ],
  "textContent": "Hi all,\n\nI built **Remoroo** , an agent for **autoresearch-style workflows** and long-running experiments.\n\nThe idea is simple: you give it a codebase and a measurable target, and it keeps iterating through the loop instead of stopping at code suggestions:\n\n**patch → run → eval → repeat**\n\nIt’s designed for agentic workflows where:\n\n  * runs take a long time\n  * results need to be measured\n  * changes should be kept or reverted based on actual outcomes\n  * you want a clear record of what was tried\n\n\n\nYou can try it with:\n\n\n    pip install remoroo\n    # or\n    uv tool install remoroo\n\n\nThen inside a toy repo:\n\n\n    remoroo run --local --engine venv\n\n\nand follow the prompts.\n\nLinks:\n\n  * Website: http://remoroo.com/\n\n\n\nIf you’re interested in a paid version, sign up first, then email **adham@remoroo.com** with the subject line **Remoroo autoresearch discount** and I’ll share a **10% off** coupon.\n\nWould love feedback from people here, especially on where this could be most useful in real HF-style workflows.",
  "title": "Built Remoroo — an agent for long-running autoresearch workflows"
}