# FutureSmart AI Demos

> Search evidence-backed AI-tool reviews, rankings, use cases, comparisons & toolkits (read-only).

- name: com.aidemos/catalogue
- version: 2.0.0
- connection class: R0 (autonomous)
- trust: 80/100 flags: no-repository
- quality: 79/100 (good)
- owner: not claimed
- descriptor: https://api.protogrid.dev/v1/servers/com.aidemos%2Fcatalogue
- tools: https://api.protogrid.dev/v1/servers/com.aidemos%2Fcatalogue/tools

**An agent can connect right now, no human step.** Remote endpoint, no authentication, reachable on the last probe.

## Connection (mcpServers)

```json
{
  "mcpServers": {
    "catalogue": {
      "type": "http",
      "url": "https://mcp.aidemos.com/mcp"
    }
  }
}
```


## Trust

- hygiene: 80
- liveness: 100
- freshness: 85
- provenance: 55
- drivers: +dns-namespace +website +reachable +uptime-99% ~updated-89d-ago -no-repository
- Derived from observable signals (official registry feed and our own credential-free probes); no code audit performed.

## Quality 79/100 (good)

### protocol: 75 (weight 25)

- warn `protocol.modern`: newest supported version is 2025-06-18; 2026-07-28 not supported, older versions are deprecated until 2027-07-28
- n/a `protocol.stateless`: only applies to 2026-07-28 servers
- pass `protocol.transport`: streamable HTTP
- n/a `protocol.list_ttl`: only applies to 2026-07-28 servers

### authorization: n/a (weight 15)

- n/a `auth.prm`, `auth.as_metadata`, `auth.cimd`: no remote uses OAuth
- n/a `auth.secret_in_url`: no templated URL

### tool hygiene: 71 (weight 30)

- pass `tools.descriptions`: every tool has a description
- pass `tools.description_length`: descriptions are concise
- pass `tools.schemas`: every tool has an object input schema
- fail `tools.annotations`: no tool declares readOnlyHint or destructiveHint, so clients cannot tell safe calls from risky ones
- fail `tools.directory_hints`: no tool declares all four of readOnlyHint, destructiveHint, idempotentHint and openWorldHint; missing: rankings_for_tool (readOnlyHint, destructiveHint, idempotentHint, openWorldHint); get_persona (readOnlyHint, destructiveHint, idempotentHint, openWorldHint); list_use_cases (readOnlyHint, destructiveHint, idempotentHint, openWorldHint); …
- pass `tools.token_cost`: about 3,045 tokens to load every tool
- pass `tools.api_dump`: tools are not a one-to-one API dump

### stability: 100 (weight 15)

- pass `stability.changes`: no tool changes in 30 days
- pass `stability.rug_pull`: no tool changed its meaning under the same name

### dependencies: n/a (weight 15)

- n/a `deps.known_vulns`, `deps.mcp_sdk_version`, `deps.resolvable`: no npm or PyPI package

- tools: 17, about 3,045 tokens to load them all
- tool set hash: 6fd4e69aacf67d69aceb63abcda592ec
- badge: [![protogrid quality](https://protogrid.dev/badge/com.aidemos/catalogue.svg)](https://protogrid.dev/servers/com.aidemos/catalogue)
- Checks run on what our credential-free, read-only probes observe; tools are never called and no code audit is performed.

### Tool changes

_No tool definition changes recorded._

## Remotes

- https://mcp.aidemos.com/mcp (streamable-http, auth none, reachable true, uptime30d 0.9915254)

## Packages

_none_

## Tools (17)

- `answer`: ANSWER a buyer question in ONE call: which tool is best at a specific capability, with proof. Returns the resolved verdict our testing team's evidence supports — a named winner FOR THE ASKED CRITERION, every tested tool ranked with a comparable score /5, the CONDITIONS each result holds under (e.g. 'clean tables yes; nested headers no'), dissenting observations preserved as openable links, the tie-break reason, and artifact proof URLs. Answers are materialized from the evidence substrate — the same question returns the same answer. Honestly refuses (coverage: not_tested) when we never tested the topic. Start HERE for any 'which tool is best at X' / 'A or B for X' question; use get_evidence for the raw cells behind it.
- `compare_tools`: Evidence-aligned comparison of two tools, honesty enforced structurally: head_to_head (cells from the SAME test input — provable same-input comparison), related_not_same_input (same dimension, different runs — flagged), and each tool's unique evidence. Built from real observation cells with artifacts, not prose.
- `get_evidence`: Query the evidence graph: observation cells (tool × test-scenario × criterion) → verdict, score, the researcher's note, and the REAL artifacts (input/output screenshots) that prove it. The ground truth behind every ranking — filter any combination of tool(s), scenario (slug, group tag, or name), criterion, verdict, or evidence state. evidence_state: "verified" = artifact-backed, "observed" = noted without artifact, "scored-only" = number only.
- `get_persona`: Given a persona slug (from list_personas), return everything tagged with it: persona, use_cases[], rankings[], compares[], toolkits[], tools[]. Mirrors a persona landing page.
- `get_ranking`: Full ranking detail as a JSON+Markdown envelope: ranked tools (rank/badge/scores), criteria, winner, breakdown (JSON) + verdicts / final take (Markdown). Every ranked tool carries `proof`: real artifact URLs from the exact evidence run this page is bound to, each with the /evidence permalink for the finding it proves — verdict and proof arrive together, no second call needed to cite evidence. `proof.artifact_count` is the true total and `proof.all_findings` is the get_evidence call that returns all of it. null if unknown. Pass `fields` to project to only the keys you need (token-efficient).
- `get_tool`: Full tool detail as a JSON+Markdown envelope: identity, pricing, per-feature scores, fit, FAQ, relationships (JSON) + our_take / in-depth review (Markdown). Includes `proof`: real artifact URLs (input/output screenshots, recordings) from the runs that tested it, each with the /evidence permalink for the finding it proves — you can cite evidence from THIS call. null if unknown. Pass `fields` to project to only the keys you need (token-efficient).
- `get_use_case`: Full use-case detail as a JSON+Markdown envelope: step guide, pros/cons, FAQ, tools used (JSON) + the full narrative guide (Markdown, full_md). null if unknown. Pass `fields` to project to only the keys you need (token-efficient).
- `list_categories`: List the category vocabulary in use with per-type counts and a source flag ("collection"|"derived"), sorted by total.
- `list_compares`: List published head-to-head comparison pages: id, title, slug, url, tool_a, tool_b, personas[], shared_use_cases[], updated_at.
- `list_personas`: List personas with published-page counts per type {use_cases, rankings, tools, compares, toolkits}. The slugs are valid input for get_persona.
- `list_rankings`: List published ranking pages ("best X"): id, title, slug, url, use_case, persona, category, tools_count, winner, tested_as_of, updated_at.
- `list_toolkits`: List published toolkit pages (curated bundles): id, title, slug, url, category.
- `list_tools`: List published AI tool pages: id, name, slug, url, domain, personas[], categories[].
- `list_use_cases`: List published use-case pages (how-to guides): id, title, slug, url, persona, category, updated_at.
- `rankings_for_tool`: Given a tool id, return the tool {id,name,slug} and every ranking it appears in [{id,title,slug,url,use_case,rank,badge}].
- `search`: Search the published catalogue across use cases, rankings, tools, comparisons, and toolkits. Returns ranked light refs [{ kind, id, title, slug, url, snippet, score, meta }] — then call get_tool / get_ranking / get_use_case for full detail. `mode`: keyword (substring), semantic (meaning, via embeddings — finds pages by what they cover), or hybrid (default, fuses both). Optional filters: type[], persona, category.
- `tools_in_ranking`: Given a ranking id, return the ranking {id,title,slug} and its ranked tools [{id,name,slug,url,rank,badge}] (badge: Best/Usable/Needs work/Unstable/Failed).