
CliniAtlas MCP server
Which servers in the official MCP registry are CliniAtlas's own, whether they answer, and what they offer. A program checked each of these, nobody copied them from a product page.
read 2026-10-20
- Registry namecom.cliniatlas/evidenceverified in the MCP registry
The registry name sits under cliniatlas.com. The registry only lets a publisher use that namespace after proving control of the domain through DNS or HTTP.
- Stateanswerschecked 2026-10-20
- Tools7
- Server calls itselfCliniAtlas Evidence Library
- Server version0.1.0
- Protocol version2025-06-18
On 2026-10-20 the server answered initialize and tools/list without a sign-in. First answer in 496 ms.
- Whether its tools do the product's job has not been judged yet.
- Checked on 2 of the 7 days it takes to judge reliability.
Once connected
- For "search PubMed and guidelines for a clinical question", the likely first call is search_medical_evidence.
- Parameters are untyped or undescribed in 7 of 7 tools: check arguments before calling.
Said by rules from the measurements above. The likely first call is a model's pick from the tool list; read its description before relying on it.
- INITE MCP scoreafter a weeka score needs 7 daily checks: 2 so far
- Answeredmeasuring2 of the 7 daily checks it takes to print a share
- Median first answer541 ms
- Space in an agent's context≈ 1,226 tokensthe definitions of all its tools (7), which an agent carries on every turn
- Tool definitionsB · 83 of 100
- Typed results7 of 7 toolsdeclared outputSchema: results come typed, read once
- Parameters without a description: in 7 of 7
What an agent can do with it
- can do · search PubMed and guidelines for a clinical question
- not sure · check each cited claim against its linked source
- not sure · compare outcome trial results across drug classes
Each of the product's use cases, put to Jev against this server's tool list: can an agent do it with these tools?
Worth a look
None of the five warning signs in these definitions.
A program checks the server once a day: initialize and tools/list, no model. Tokens are the definitions' length at about four characters a token. The grade averages six checks of every tool's name, description, schema and annotations; it says how well the tools are described, not whether they work. The score weighs what was measured and nothing else: answering (30%), doing the product's job (25%), definitions (20%), context (10%) and use (15%). A part that was not measured is left out, and a hidden instruction caps the score at 40. The state of vendors' MCP servers: method and numbers →
search_evidence_libraryFind published CliniAtlas evidence answers matching a clinical question. Call this first for a clinical question: it is instant and does not use the user's free daily answers. Returns question, summary and link for each match. The library…
get_evidence_answerFull published answer: cited paragraphs, limitations and graded sources. Use a slug from search_evidence_library. Citation markers like [1] refer to the numbered sources. Show the sources and intended_use with the answer.
lookup_pubmedCitation metadata for a PubMed ID: title, journal, year, authors, DOI and design. Abstracts are not returned; follow pubmed_url to read them.
lookup_eu_medicineCentrally authorised EU medicines by brand name or active substance. Returns the EMA summary, authorisation status and ema_url, the EMA page whose Product information section holds the SmPC. Nationally authorised medicines are not in this…
search_medical_evidenceLive search of PubMed, Europe PMC and clinical guidelines. Returns up to eight sources for a clinical or research question, each with its study type, evidence level and link, usually in about 20 seconds. It does not write an answer: cite…
ask_cliniatlasFull CliniAtlas answer to a clinical or research question, with numbered citations. Searches the literature, guidelines and EMA documents and writes a cited answer, the same answer cliniatlas.com gives. Uncited parts are reported through…
get_cliniatlas_answerCollect the answer for a job_id returned by ask_cliniatlas. Waits up to about 40 seconds for the result. If status is still "running", call it again with the same job_id.
How to connect
Add the server to your MCP client config. Claude Desktop, Cursor and most other clients read this shape:
{
"mcpServers": {
"evidence": {
"type": "http",
"url": "https://ydjbww4h7sfqc3sf2wcf26mfbq0bwvfu.lambda-url.eu-central-1.on.aws/public/mcp"
}
}
}Add it
Claude Code, in a terminal
claude mcp add --transport http cliniatlas https://ydjbww4h7sfqc3sf2wcf26mfbq0bwvfu.lambda-url.eu-central-1.on.aws/public/mcpclaude.ai, ChatGPT and other clients: add https://ydjbww4h7sfqc3sf2wcf26mfbq0bwvfu.lambda-url.eu-central-1.on.aws/public/mcp as a custom connector in the settings.
Each line is something a program found when it went looking: the registry entry, the server's own answer, a spec that parses. A missing line means we did not find it, not that it does not exist.
Cookies
Before we count this visit
Analytics and ad measurement tell us which pages and which of our ads do their job: Google Analytics and Google Ads, the Meta pixel, Yandex Metrica, and one first-party id of our own. Nothing is stored until you answer, and declining changes nothing about the site.