Checks RAG citations against their sources (CrossSource method), plus claim checks. Read-only, free.
Maintain this server? Claim it
Call this server's tools from your browser. Nothing to install.
Requests go through mcptry because browsers can't call most MCP servers directly. Header values are sent to the server for that request only and are never stored or logged.
Tool list from our last check. Connect live to refresh it.
Check whether each inline citation in an answer is supported by the source it points to, using CrossSource's method and failure taxonomy (supported, right_doc_wrong_passage, wrong_document, unsupported_claim). Cites to ids not in `sources` and citations where another passage clearly covers the claim are decided deterministically. Everything else returns needs_judgment with the CrossSource judge prompt to apply. Set use_client_model to have your client model judge those, where the client supports MCP sampling.
The answer text with inline citations like [doc_id:chunk_id] or [source_id].
Every passage the answering system had available (30 at most).
Ask the client model (MCP sampling) to judge needs_judgment items. Default false.
Return the judge prompt with each needs_judgment item. Default true.
Pick your client. Values in capitals, like YOUR_TOKEN, are yours to fill in.
Run in a terminal
claude mcp add --transport http zoebnomi https://mcp.zoebnomi.com/mcp
Check whether each inline citation in an answer is supported by the source it points to, using CrossSource's method and failure taxonomy (supported, right_doc_wrong_passage, wrong_document, unsupported_claim). Cites to ids not in `sources` and citations where another passage clearly covers the claim are decided deterministically. Everything else returns needs_judgment with the CrossSource judge prompt to apply. Set use_client_model to have your client model judge those, where the client supports MCP sampling.
| Argument | Type | Description |
|---|---|---|
| answer* | string | The answer text with inline citations like [doc_id:chunk_id] or [source_id]. |
| sources* | array | Every passage the answering system had available (30 at most). |
| use_client_model | boolean | Ask the client model (MCP sampling) to judge needs_judgment items. Default false. |
| include_judge_prompts | boolean | Return the judge prompt with each needs_judgment item. Default true. |
{
"type": "object",
"$schema": "http://json-schema.org/draft-07/schema#",
"required": [
"answer",
"sources"
],
"properties": {
"answer": {
"type": "string",
"maxLength": 30000,
"minLength": 1,
"description": "The answer text with inline citations like [doc_id:chunk_id] or [source_id]."
},
"sources": {
"type": "array",
"items": {
"type": "object",
"required": [
"id",
"text"
],
"properties": {
"id": {
"type": "string",
"maxLength": 200,
"minLength": 1,
"description": "The id used inside the brackets."
},
"text": {
"type": "string",
"maxLength": 12000,
"description": "The passage text (12,000 characters at most)."
},
"doc_id": {
"type": "string",
"maxLength": 200,
"description": "Document id; defaults to the part of `id` before \":n\"."
}
}
},
"maxItems": 30,
"description": "Every passage the answering system had available (30 at most)."
},
"use_client_model": {
"type": "boolean",
"description": "Ask the client model (MCP sampling) to judge needs_judgment items. Default false."
},
"include_judge_prompts": {
"type": "boolean",
"description": "Return the judge prompt with each needs_judgment item. Default true."
}
}
}Map a job description to Zoeb Nomi's public work (zoebnomi.com) with quoted evidence, and list requirements with no public evidence as gaps. Never answers compensation, availability, location or work authorization.
| Argument | Type | Description |
|---|---|---|
| job_description* | string | The job description text. |
{
"type": "object",
"$schema": "http://json-schema.org/draft-07/schema#",
"required": [
"job_description"
],
"properties": {
"job_description": {
"type": "string",
"maxLength": 20000,
"minLength": 20,
"description": "The job description text."
}
}
}Check a claim or number about Zoeb Nomi against what zoebnomi.com states. Returns found, number_mismatch or not_found, with the site quote, source URL, evidence link, n and date where the site gives them. Use the returned quote, not your own paraphrase. A word lookup: it cannot see negation, so read the quote.
| Argument | Type | Description |
|---|---|---|
| claim* | string | The claim to check, e.g. "CrossSource lifted citation precision to 0.994". |
{
"type": "object",
"$schema": "http://json-schema.org/draft-07/schema#",
"required": [
"claim"
],
"properties": {
"claim": {
"type": "string",
"maxLength": 1000,
"minLength": 3,
"description": "The claim to check, e.g. \"CrossSource lifted citation precision to 0.994\"."
}
}
}We read every tool's name, description and inputs the way the model will, looking for instructions hidden from you. A clean result means no warning signs, not a guarantee. Scanned Nov 20, 2026.