# Speakeroo speaking coach

> AI speaking coach: feedback on speeches from text or audio, plus topics, pronunciation and rubrics.

- name: app.speakeroo/speaking-coach
- version: 1.0.0
- connection class: R0 (autonomous)
- trust: 82/100 flags: no-repository
- quality: 87/100 (new)
- owner: not claimed
- descriptor: https://api.protogrid.dev/v1/servers/app.speakeroo%2Fspeaking-coach
- tools: https://api.protogrid.dev/v1/servers/app.speakeroo%2Fspeaking-coach/tools

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

## Connection (mcpServers)

```json
{
  "mcpServers": {
    "speaking-coach": {
      "type": "http",
      "url": "https://mcp.speakeroo.app/mcp"
    }
  }
}
```


## Trust

- hygiene: 80
- liveness: 98
- freshness: 100
- provenance: 55
- drivers: +dns-namespace +website +reachable +uptime-95% +updated-5d-ago -no-repository
- Derived from observable signals (official registry feed and our own credential-free probes); no code audit performed.

## Quality 87/100 (new)

### protocol: 88 (weight 25)

- pass `protocol.modern`: supports 2026-07-28 (server/discover)
- pass `protocol.stateless`: answers without a session
- pass `protocol.transport`: streamable HTTP
- warn `protocol.list_ttl`: tools/list declares ttlMs 0, so clients re-fetch it on every use

### 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: 86 (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
- pass `tools.annotations`: 7 of 7 tools declare readOnlyHint or destructiveHint
- fail `tools.directory_hints`: no tool declares all four of readOnlyHint, destructiveHint, idempotentHint and openWorldHint; missing: analyze_transcript (destructiveHint, idempotentHint); refine_transcript (destructiveHint, idempotentHint); get_speech_topics (destructiveHint, idempotentHint); …
- pass `tools.token_cost`: about 1,439 tokens to load every tool
- pass `tools.api_dump`: tools are not a one-to-one API dump

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

- n/a `stability.changes`, `stability.rug_pull`: tool history recorded for 4 of the 7 days needed

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

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

- tools: 7, about 1,439 tokens to load them all
- tool set hash: 8acd942d14139081bb0d5001bfd69400
- badge: [![protogrid quality](https://protogrid.dev/badge/app.speakeroo/speaking-coach.svg)](https://protogrid.dev/servers/app.speakeroo/speaking-coach)
- 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.speakeroo.app/mcp (streamable-http, auth none, reachable true, uptime30d 0.94736844)

## Packages

_none_

## Tools (7)

- `analyze_speech`: Analyze a spoken recording: it is transcribed, then returns everything analyze_transcript does plus delivery metrics only audio can give: words per minute, pace, pauses per minute, filler rate, pitch (median and range) and vocal expression, and pacing, pronunciation and engagement scores. Limits: up to 120 seconds and 10 MB; mp3, m4a/mp4, wav, webm, ogg or flac; at least 30 spoken words. Pass audioUrl (public https) or audioBase64. Free: 1 analysis a day per network; a try that fails on our side does not count, up to 3 tries a day. Audio is deleted once the response is sent.
- `analyze_transcript`: Coach a speech from its text. Returns weak and filler words with firmer alternatives, rhetorical devices, delivery and rhetoric feedback, grammar/vocabulary/coherence scores (1-10) on the Speakeroo rubric, and word-level issues with the excerpt and a fix. Input: 30-1500 words. Text cannot show pace, pauses, pitch or pronunciation; use analyze_speech with a recording for those. Free, 5 text analyses a day per network, shared with refine_transcript.
- `get_speaking_rubric`: A four-level speaking rubric teachers use for speaking homework: fluency, pronunciation, vocabulary range and task completion, with a descriptor for every level. Links to the page with a CSV download. analyze_transcript and analyze_speech score on Speakeroo's own 1-10 rubric.
- `get_speech_topics`: Random speech topics to practise with: impromptu, persuasive, informative, presentation, IELTS Speaking Part 1 questions or Part 2 cue cards. Each result links to the full list on speakeroo.app, where every topic opens a timed practice session.
- `how_to_pronounce`: How to pronounce an English word Speakeroo has a guide for: IPA, a plain respelling, syllables with the stressed one, sound tips, a note on the usual mistake and two example sentences. Links to the guide, where you can record yourself and get a pronunciation check.
- `list_filler_words`: The common filler words in spoken English grouped by what they do, and the steps to stop using them. Links to the guide on speakeroo.app, which has a free filler word tracker.
- `refine_transcript`: Rewrite a speech into a clear structure (PREP, STAR, problem-solution, rule of three and others) chosen to fit it. Returns the rewritten sections, the structure with a 1-10 score for how well the original already followed it, a summary of what changed and one delivery tip. Input: 30-1500 words. Shares the daily text allowance with analyze_transcript.