# Uttera

> Audio for your agent: transcribe, speak, translate, summarise, plus sound effects and music.

- name: ai.uttera/uttera
- version: 1.0.0
- connection class: R2 (oauth_consent_once)
- trust: 82/100 flags: no-repository
- quality: 85/100 (good)
- owner: not claimed
- descriptor: https://api.protogrid.dev/v1/servers/ai.uttera%2Futtera
- tools: https://api.protogrid.dev/v1/servers/ai.uttera%2Futtera/tools

**An agent can connect after a one-time OAuth consent by a human.** Remote endpoint behind OAuth 2. Tokens stay with the agent; the registry only exposes the authorization metadata.

## Connection (mcpServers)

```json
{
  "mcpServers": {
    "uttera": {
      "type": "http",
      "url": "https://mcp.uttera.ai/mcp"
    }
  }
}
```


## Trust

- hygiene: 80
- liveness: 99
- freshness: 100
- provenance: 55
- drivers: +dns-namespace +website +reachable +uptime-97% +updated-27d-ago -no-repository
- Derived from observable signals (official registry feed and our own credential-free probes); no code audit performed.

## Quality 85/100 (good)

### protocol: n/a (weight 25)

- n/a `protocol.modern`, `protocol.stateless`, `protocol.transport`, `protocol.list_ttl`: no remote answered a protocol handshake (not probed yet, unreachable, or local-only)

### authorization: 83 (weight 15)

- pass `auth.prm`: Protected Resource Metadata (RFC 9728) published
- pass `auth.as_metadata`: authorization server metadata reachable
- warn `auth.cimd`: only Dynamic Client Registration, deprecated in 2026-07-28 in favour of CIMD
- 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`: 9 of 9 tools declare readOnlyHint or destructiveHint
- fail `tools.directory_hints`: no tool declares all four of readOnlyHint, destructiveHint, idempotentHint and openWorldHint; missing: generate_sound_effect (idempotentHint); transcribe_audio (idempotentHint); summarize_recording (idempotentHint); …
- pass `tools.token_cost`: about 1,097 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 not recorded yet

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

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

- tools: 9, about 1,097 tokens to load them all
- badge: [![protogrid quality](https://protogrid.dev/badge/ai.uttera/uttera.svg)](https://protogrid.dev/servers/ai.uttera/uttera)
- 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.uttera.ai/mcp (streamable-http, auth oauth2, reachable true, uptime30d 0.9736842)

## Packages

_none_

## Tools (9)

- `generate_music`: A finished piece of music from a description. Up to 380 s.

    Needs a paid plan (Startup and up). ⚠ BOTH controls cost: the price is
    length multiplied by steps, so 380 s at 128 steps is about eight times a
    60 s at 32. Naming style, instruments and BPM helps far more than asking
    for «something upbeat».
    
- `generate_sound_effect`: One sound effect from a description: a door, rain, footsteps. Up to 30 s.

    Needs a paid plan (Startup and up). The description works better in English.
    A fixed `seed` reproduces the same sound exactly.
    
- `identify_speakers`: Who speaks and when. Returns the segments with their speaker and times.
- `last_call_cost`: What the previous call actually cost, in credits, broken down.

    Worth checking before running something many times: it is the honest way
    to tell the user what a batch is going to cost them.
    
- `list_voices`: The voice catalogue, with the language each one speaks.
- `speak_text`: Read text aloud. Returns audio. `voice` defaults to a Spanish one;
    call list_voices to see the catalogue.

    ⚠ Line breaks are billed: each one inserts a pause of about 1.3 s. Strip
    them from formatted text before sending it.
    
- `summarize_recording`: Summarise a long recording: summary, full transcript, tone, speaker
    profile and who spoke when — all in one request and one upload.
- `transcribe_audio`: Transcribe a recording. `audio` is an https URL or base64 audio.

    Use this for short recordings when you want the literal text. For anything
    past a few minutes prefer summarize_recording: it returns the transcript
    too, plus who spoke and the tone, in ONE upload instead of two.
    
- `translate_recording`: Translate a recording into another language: text and, where the plan
    allows it, voice. `target_language` is a code such as en, es, fr, de.