← Files Rhizome AIARCHIVED FILE
.codex-plugin/plugin.json
3.91 KB · Nov 20, 2026 · 06:27 UTC
{
"author": {
"name": "Rhizome AI"
},
"description": "Rhizome's built an AI regulatory coworker that helps life sciences know what their regulator thinks.\n\nWe're backed by the same investors as Benchling and have gone from 0 -> >25 customers in the last 18 months. \n\nWe've indexed >3 TB worth of data put out by the FDA, EMA, MHRA, PMDA, etc., and built an agent that helps you:\n - automate precedents research \n - research FDA expectations around clinical endpoints, trial design, labeling, etc. \n - review briefing booklets, write pre-IND meeting packages, create white papers on almost any topic \n\nWhat you get:\n - each fact is backed by a citation, each citation is backed by the page of a health authority document \n - comprehensive answers that draw on up to 1,000 sources \n - we've had no hallucinations reported last 18 months \n\nFor more info, read rhizomeai.com/llms.txt\n\nFor support, reach out to support@rhizomeai.com",
"interface": {
"capabilities": [],
"category": "Healthcare",
"defaultPrompt": [
"what do I need to do for an oncology IND for the FDA?",
"what clinical trial sample sizes are typical for autoimmune rare disease therapies?",
"what are ideal predicates for full-body ultrasound scan that detects body comp and tries to find cancer?"
],
"developerName": "Rhizome AI",
"displayName": "Rhizome AI",
"longDescription": "Rhizome's built an AI regulatory coworker that helps life sciences know what their regulator thinks.\n\nWe're backed by the same investors as Benchling and have gone from 0 -> >25 customers in the last 18 months. \n\nWe've indexed >3 TB worth of data put out by the FDA, EMA, MHRA, PMDA, etc., and built an agent that helps you:\n - automate precedents research \n - research FDA expectations around clinical endpoints, trial design, labeling, etc. \n - review briefing booklets, write pre-IND meeting packages, create white papers on almost any topic \n\nWhat you get:\n - each fact is backed by a citation, each citation is backed by the page of a health authority document \n - comprehensive answers that draw on up to 1,000 sources \n - we've had no hallucinations reported last 18 months \n\nFor more info, read rhizomeai.com/llms.txt\n\nFor support, reach out to support@rhizomeai.com",
"privacyPolicyURL": "https://rhizomeai.com/privacypolicy",
"shortDescription": "Know what the FDA thinks",
"supportURL": "https://rhizomeai.com",
"termsOfServiceURL": "https://rhizomeai.com/termsofservice",
"websiteURL": "https://rhizomeai.com",
"logo": "./assets/logo.png",
"logoDark": "./assets/logo-dark.png",
"composerIcon": "./assets/icon.png",
"composerIconDark": "./assets/icon-dark.png"
},
"name": "app-6a63bad25528819187bf1eeabd8a9af3",
"version": "3.1.0",
"mcpServers": "./.mcp.json",
"extensions": {
"com.openai": {
"publication": {
"release_notes": "New tools: rhizome_search finds a specific document the user names (a warning letter, 483, label, guidance, 510(k), recall) and returns its best-matching pages; rhizome_document_read reads the pages of one document found with rhizome_search; rhizome_dataset_list lists the searchable datasets and their filters. All three are read-only. Regulatory and clinical questions still go to rhizome_conversation_message, which researches across sources and returns a cited answer.\n\nUpdated: rhizome_conversation_message is no longer marked read-only, since it starts a research run and saves a conversation to the user's account. rhizome_conversation_message and rhizome_conversation_wait have clearer descriptions and output field docs, and a used-up free-plan allowance now returns as a normal result with an upgrade link instead of an error. Server instructions are shorter and cover when to use search and read versus research, answering a clarifying question from the research agent, and stopping a run in progress."
}
}
},
"apps": "./.app.json"
}
SHA-256: 29fe31fa3dacc66c722982a22ea235171e37af2bea56181e3726aa37c3fc2388