Skip to main content
Glama

aw.speech.whisper_encode_step

Prepare the exact original Tiny or Base encoder in separate bounded calls: two for Tiny, three for Base. Begin with the accepted WAV source and no continuation. While status is encoder-pending, supply its opaque continuation with the same session, source and model and a new operationId. Retry the same operationId with identical arguments only; no automatic model retry. After encoder-prepared, choose decoderRoutes.fileOnly for a whole file without replacing the current transcript, or decoderRoutes.currentTranscript for a whole file plus shared current segments, export and Clear. Both use the same prepared handle, session, source and model; start with null decoder continuation. Never switch tools with a pending decoder continuation. Intermediate activations stay in the encrypted session; no upload, download or transcript is produced by this tool.

Tool call template

Replace each placeholder with a value matching the input schema. Only required parameters are included; add optional parameters as needed.

{
  "id": 1,
  "jsonrpc": "2.0",
  "method": "tools/call",
  "params": {
    "arguments": {
      "session": "<session>",
      "operationId": "<operationId>",
      "source": "<source>"
    },
    "name": "aw.speech.whisper_encode_step"
  }
}

Input Schema

NameRequiredDescriptionDefault
modelNoonnx-community/whisper-tiny
sourceYesAn immutable input or result handle belonging to this session; never a URL/path/base64.
sessionYesPrivate short-lived session capability returned by aw.session.open. Never publish or log it.
operationIdYes
continuationNo

Schema Changelog

Changes observed during successful MCP inspections.

  1. First observed

TDQS

A4.7/5.0
Behavior5/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

Annotations only say readOnlyHint=false, destructiveHint=false, openWorldHint=false; the description adds substantive behavior: intermediate activations stay in the encrypted session, no upload/download/transcript is produced, retries are idempotent per operationId, and the fileOnly route does not replace the current transcript. It also discloses the multi-call, resumable nature of the operation, which annotations cannot convey.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness4/5

Is the description appropriately sized, front-loaded, and free of redundancy?

Front-loaded with the core action and call count, and every sentence carries protocol information that an agent needs. However, it is written as a dense run-on with clauses like 'a whole file plus shared current segments, export and Clear' that are hard to parse, and the forward reference to decoderRoutes would read more clearly if it named the sibling tool.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness4/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

For a stateful, multi-call tool with no output schema and only 40% schema coverage, the description covers the full loop: first call, pending state, retry rules, prepared state, and hand-off to the decoder route. It stops short of saying where 'status' is returned or read from, and its decoderRoutes/null-continuation phrasing is imprecise about tool boundaries.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters4/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema description coverage is only 40%, and the description compensates well: it explains that source must be the accepted WAV, session/source/model must stay identical across stages, operationId is new per continuation but identical on retry, and continuation is the opaque resumption token. Two defects hold it below 5: it instructs 'start with null decoder continuation' although the schema types continuation as a pattern string with additionalProperties=false, and it references 'decoderRoutes.fileOnly/currentTranscript' as if selectable here, which is not a property of this schema.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose5/5

Does the description clearly state what the tool does and how it differs from similar tools?

States a specific verb and resource: 'Prepare the exact original Tiny or Base encoder in separate bounded calls', and even quantifies the work (two calls for Tiny, three for Base). It distinguishes this from the decoder stage by explicitly deferring decoderRoutes selection until after 'encoder-prepared', so an agent can separate it from aw.speech.whisper_decode_step without opening either schema.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines5/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

Gives explicit sequencing: begin with the accepted WAV source and no continuation; while status is encoder-pending, resend the continuation with a new operationId; only after encoder-prepared choose a decoder route. It also states a prohibition ('Never switch tools with a pending decoder continuation') and a retry rule (same operationId with identical arguments only, no automatic model retry). This is the when/when-not/alternative guidance the dimension rewards.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

Try in Browser

Glama MCP Gateway

Add one secure layer between your agents and this server.

Resources