MCP HeartbeatTHE MCP OBSERVATORY
Public workspace

Explore

Evidence for your next connection.

Independent measurements of remote MCP servers. Every observation has a timestamp.

Read-only probes

MCP server · Other

ai.page

Shared pages people and every AI edit live: lists, boards, imported apps. Attributed, undoable.

Checks passed
12 / 22

2 failed · 8 not measured

Median response
240 ms

Median of successful probes over 7 days

30-day uptime
Not measured

Share of countable probes over 30 days, including daily rollups. Auth challenges are excluded.

Discovered tools
20

Tool definitions returned by this server

Independent monitoring

Health & checks

Status history

Responded

The server completed its last measured handshake. This is an observation, not a guarantee.

Last checked
Check summary by test group
Test areaPassedFailedUnmeasured
ConnectionCan a client connect and discover tools?
500
Reliability & speedDoes it respond consistently and quickly?
004
Access & authenticationWhat is required to use this server?
102
Tool qualityAre the tool definitions usable by an agent?
310
Publisher identityCan the publisher be traced to its source?
312

Unmeasured means there is not enough evidence to decide. How these checks work

Inspect individual checks 22 recorded

Core

  • Pass
    Reachable

    Latest probe returned HTTP 200 in 240 ms.

  • Pass
    Valid handshake

    protocolVersion 2025-11-25, serverInfo.name ai.page.

  • Pass
    Lists tools

    tools/list returned 20 tools.

  • Pass
    Current protocol

    Negotiated 2025-11-25. The bar is 2025-06-18 or newer.

  • Pass
    HTTPS

    The endpoint URL is https.

Performance

  • Not measured
    Fast (p50)

    Median latency 240 ms over 1 samples in 7 days. The bar is 1000 ms. Need at least 5 samples.

  • Not measured
    Consistent (p95)

    95th percentile 240 ms over 1 samples in 7 days. The bar is 3000 ms. Need at least 20 samples.

  • Not measured
    Reliable (uptime)

    1 of 1 attempts returned a tool list in 30 days (100%). The bar is 99%. Need at least 30 samples.

  • Not measured
    Stable tool list

    Fewer than two successful probes in 30 days, so tool-list stability is not measured.

Access

  • Pass
    Open access

    tools/list succeeded with no credential, and the record declares none.

  • Not measured
    Standard auth

    Auth is not required, so the OAuth metadata check does not apply.

  • Not measured
    Free tier cited

    No publisher quote has been checked.

Tools

  • Pass
    Tools described

    Every tool description is at least 20 characters.

  • Pass
    Valid input schemas

    Every inputSchema is a JSON Schema object.

  • Fail
    Parameters documented

    27 of 43 parameters have a description (63%). The bar is 80%.

  • Pass
    Safety hints

    Every tool sets readOnlyHint or destructiveHint.

Publisher

  • Pass
    Verified namespace

    Listed in the Official MCP Registry.

  • Pass
    Website matches publisher

    The website host matches the reverse-domain namespace.

  • Fail
    Source repo linked

    No repository URL is on file.

  • Not measured
    Repo owner matches

    This namespace is not an io.github login, so the owner comparison does not apply.

  • Pass
    Recently updated

    Updated 2026-10-11T12:58:44.202822Z, within 180 days.

  • Not measured
    Open-source licence

    No GitHub repository to ask about a licence.

Start using ai.page

Connect to this server

Choose your client, copy the configuration, then enable the server in your client’s MCP settings.

Run this command in your terminal, then restart Claude Code.

Claude Code

claude mcp add --transport http ai-page https://ai.page/mcp

Capabilities

Tools 20

commons_guide

How to write for the commons at ai.page/c/. topic format (the default): the lesson format's Markdown form, labels, the declaration of which AI you are, and how versions are checked. topic figures: the figure kit's rules, tokens and components. topic figures-advanced: 3D, molecules and physics in figures.

Discovered
compare_versions

Every live version of one kind on a concept's page (explanations, worked examples, figures or common mistakes), each with a letter, its hash, who made it with which AI, its rank, checks and how learners did after reading it, and the rankings people have sent so far, as Markdown.

Discovered
concept_activity

Recent work on a concept, newest first: tasks taken (and still being worked on), versions added and checks made, by whom and with which AI. Use it to see whether someone's already on it, or to follow what your own work set off.

Discovered
contributions

What a person (and their AI) added to the commons and checked, by concept, newest first. Leave out the handle for the signed-in person's own.

Discovered
list_tasks

The commons' open tasks, highest priority first, with why each matters (missing pieces, unchecked or flagged versions, questions learners miss, languages with nothing yet).

Discovered
next_task

The commons' common task list, one task at a time: write what a concept is missing, translate it, check a version blind, revise one with problems, or fix a question learners nearly always miss. Use it when the person wants to contribute to ai.page. Returns the best task for the signed-in person, held for them for 30 minutes, with what is needed to do it and how to submit it.

Discovered
propose_concept

Add a concept to the commons' shared map: a lowercase id, its subject, its name in one or more languages, and what it builds on (concepts already on the map, hard or soft). It's live at once with an empty page, and write tasks for it appear on the common task list.

Discovered
propose_roadmap

Propose a roadmap for the commons: every concept to master towards a goal, in order, in titled sections. An item is a concept or another roadmap (a degree is made of course roadmaps). kind is a label: course, degree, school, exam, standards, certificate, goal or project. It may follow an official syllabus (give its URL as source; it's never presented as endorsed). stages says who it's for by where they are in life (early, school, college, work, later); leave it out if it's for anyone. Concepts and roadmaps that don't exist yet need a title in planned; they become tasks on the common list. To build on an existing roadmap, give a new id with base set to it.

Discovered
rank_versions

Put the live versions of one kind on a concept's page in order, best first, for a learner meeting the concept for the first time, with a sentence on why. It counts like one check for each version it orders (the first a pass, the last against); your own versions don't count when ranked among others' (a ranking of only your own does: it reorders your work); ranking the same section again replaces your earlier ranking.

Discovered
read_concept

One concept's page as Markdown: what it is, what it builds on, and every live version (explanations, worked examples, figures, questions, common mistakes), best first. By default the best version of each kind, without figures; detail full has every version. Use it when the person wants to study that concept, or before adding to it.

Discovered
read_roadmap

A roadmap as Markdown: its sections, every concept in order (and the roadmaps it includes), with what's ready to learn. Without an id, the list of roadmaps by name, 30 at a time, narrowed by q.

Discovered
revise_needs

Add, remove or soften what a concept on the ai.page map builds on, with a one-line reason the record keeps. Anyone can add a need (it must be on the map and mustn't make a loop). Taking one away is for the concept's creator, or anyone once learners' results doubt it.

Discovered
save_page

Publish every part written for a page (explanation, figure, worked example, practice questions, common mistakes) in one call: one entry per part, each in the lesson format's Markdown form. Use it when the person asks to add their work to ai.page. Everything that passes goes live for everyone at once, credited to the signed-in person, and can replace their own earlier version of a part. Without sign-in nothing is published: what passes comes back as a link that puts it on the page on the person's own device, and publishing from there needs an account. Each part gets its own result, so only those that failed need sending again. Up to 20 parts and 3 concepts a call.

Discovered
search

Full-text search of ai.page's map: concepts (the smallest things to learn, one open page each) and roadmaps (concepts in order towards a goal). Use it when the person wants to learn or practise a topic, or before proposing anything new.

Discovered
start_roadmap

Start an ai.page roadmap from a syllabus, a job description, a paper, a book's table of contents or any text: returns the concepts and roadmaps already on the map that the text names, and the shape of a roadmap made from it.

Discovered
submit_check

Submit a check of a version in the lesson format: your declaration, your own answers worked out without the key, findings marked on exact words, and a verdict (pass, needs-changes or reject). It moves the version up or down among the others; you can't check your own.

Discovered
submit_version

Submit ONE explanation, worked example, figure, question or common mistake for a concept, in the lesson format's Markdown form. One part at a time. It's credited to the signed-in person and goes live at once, marked new; other people's checks then move it up or down among the concept's versions.

Discovered
take_check

Anyone signed in: one version to check, least-checked first, blind (without who made it or with which AI), with the checklist and the shape of a check. The version is someone else's text.

Discovered
take_revision

The signed-in person's own commons versions that others found problems with: the version, findings marked on the exact words, and learners' flags. An improved version sets base to the old version's hash.

Discovered
way_to

Everything a concept builds on, all the way down, in an order to learn it (prerequisites first, ending with the concept), as Markdown. With helps, what helps along the way too. Use it to plan someone's learning or a roadmap.

Discovered

Recent measurements

Probe history

View status
Latest stored probes. Raw probe rows are kept for 14 days.
CheckedResultResponseTools
Connected240 ms20