beta-test-2431.replit.app
Search ClinicalTrials.gov, compare studies and prepare questions for research teams. A WebMCP application: browser AI agents can drive the same interface you see. Not medical advice.
First seen 2026-11-05 · Last seen 2026-11-05 · Source: webmcp.com
Answer 2 · Act 8 · Transact 0 · Score: Not yet scored
Tool inspector
record_prescreening_responses
Record what the person told you about specific criteria in the open pre-screening session, together with your own cautious comparison of each answer against the criterion you quoted. TrialBridge does not interpret criteria itself, so your comparison is what gets stored and displayed — always beside the verbatim criterion and always labelled as an agent-assisted preliminary comparison. Record only answers the person actually gave; never infer, assume or fill in an answer on their behalf. Use 'appears_consistent' when their answer lines up with the criterion, 'potential_conflict' when it may not and is worth raising with the study team, and 'unresolved' whenever they did not know, declined, or skipped. Keep each explanation factual and specific to the one criterion. Do not state or imply eligibility, do not give any percentage, score or count of criteria met, do not advise on treatment, and never discourage anyone from contacting the study team.
Input schema
| Property | Type | Required | Description |
|---|---|---|---|
nctId |
string | yes | The study being pre-screened. Must match the session opened by start_trial_prescreening. |
responses |
array | yes | Up to 10 answered criteria per call. |
Raw JSON schema
{
"type": "object",
"properties": {
"nctId": {
"type": "string",
"pattern": "^NCT\\d{8}$",
"description": "The study being pre-screened. Must match the session opened by start_trial_prescreening."
},
"responses": {
"type": "array",
"minItems": 1,
"maxItems": 10,
"description": "Up to 10 answered criteria per call.",
"items": {
"type": "object",
"properties": {
"criterionId": {
"type": "string",
"description": "The criterionId exactly as returned by start_trial_prescreening. It encodes the NCT id, so it cannot be used on another study."
},
"questionAsked": {
"type": "string",
"maxLength": 400,
"description": "The plain-language question you actually put to the person."
},
"patientAnswer": {
"type": [
"string",
"number",
"boolean",
"null"
],
"description": "What the person answered, in their own terms. Use null when they did not answer."
},
"answerType": {
"type": "string",
"enum": [
"text",
"number",
"boolean",
"unknown",
"skipped"
],
"description": "Use 'unknown' when they did not know and 'skipped' when they declined or preferred not to say."
},
"comparison": {
"type": "string",
"enum": [
"appears_consistent",
"potential_conflict",
"unresolved"
],
"description": "Your cautious reading of that answer against this one criterion. Must be 'unresolved' whenever the answer is unknown, skipped or absent."
},
"explanation": {
"type": "string",
"maxLength": 600,
"description": "One or two sentences saying why, referring only to this criterion and what the person told you."
}
},
"required": [
"criterionId",
"questionAsked",
"patientAnswer",
"answerType",
"comparison",
"explanation"
],
"additionalProperties": false
}
}
},
"required": [
"nctId",
"responses"
],
"additionalProperties": false
}
Similar websites
pocketgull.app
Pocket Gull is a real-time clinical AI consult engine powered by Google Gemini. Features 10D multi-paradigm strategy, FHIR R4 export & 3D spatial anatomy.
webmcp-six-lac.vercel.app
AI Navigation Layer: Model Context Protocol (MCP) Semantic DOM integration for intelligent, safe, and observable web navigation.
ghqsandbox.net
Interactive math and physics tools an AI agent can drive directly through WebMCP.
app.formative.com
Formative classroom assessment and learning platform