Skip to content

speech_to_text

Transcribe audio to text with word-level timings via Seed Speech ASR.

Accepts an HTTPS URL, Base64 data, or a local file path (stdio only), up to 200 MB. Requires background execution (ark_job_submit).

Requirement Value
Registration Requires BYTEPLUS_SEED_SPEECH_API_KEY.
Execution Background required
Required permission seed:asr:transcribe

This site lists the full configured capability surface, not the tools enabled on your particular server. Use model discovery and job discovery to check your connection. Credentials are server configuration, never tool arguments.

These are supported families and built-in model or service IDs for this site revision, not provider account access or a guarantee that every row is enabled on your server. Check model discovery for configured bindings and service availability.

Model Model or service ID Availability Capabilities
Seed Speech ASR Service-managed; no selectable model ID BYTEPLUS_SEED_SPEECH_API_KEY required. Background transcription with utterance/word timings when returned by the provider; asynchronous provider submission and polling handled inside the Ark job.

This tool has no per-call model argument. It uses the Seed Speech ASR service rather than a selectable ModelArk model.

Ask your connected AI client to use speech_to_text for the operation you want. The client constructs the arguments and handles tool calls. Review its proposed operation before approving anything billable. For a non-billable introduction, start with discovery without generation.

Submit through ark_job_submit with tool_name and the complete target arguments, including its inner input object. Poll ark_job_get using the returned Ark job_id, respecting poll_after_ms. Inspect both terminal status and result.is_error; the original output is in result.structured_content. Do not call this tool in the foreground.

Match the input fields below and the full JSON Schema. See the worked examples for this capability and Background Execution for job/result handling.

The MCP arguments object wraps these fields as {"input": {...}}.

Field Type Required Description and constraints
audio AsrAudioInput Yes Audio source to transcribe.
options AsrRequestOptions or null No Optional transcription feature toggles.; Default: null
Field Type Required Description and constraints
log_id string or null No Provider-side log ID for troubleshooting.; Default: null
result TranscriptionResult Yes Full transcription result with text, utterances, and timing.

Field types named above refer to these shared definitions. Optional/null fields, enum choices and bounds are retained in the complete schemas below. Schema validation is followed by runtime capability and security checks.

AsrAudioInput
Field Type Required Description and constraints
audio_data string or null No Base64 audio bytes (max 200 MB decoded). Exactly one audio source.; Default: null
audio_file_path string or null No Absolute local file path (stdio only, max 200 MB). Exactly one audio source.; Default: null
audio_format “wav”, “mp3”, “ogg”, “raw”, “flac” Yes Format of the source audio.
audio_url string or null No HTTPS URL of the audio (max 200 MB). Exactly one audio source.; Default: null
AsrRequestOptions
Field Type Required Description and constraints
enable_itn boolean or null No Inverse text normalization (e.g. spoken numbers to digits).; Default: null
enable_punc boolean or null No Add punctuation to the text.; Default: null
language string No BCP-47 language code, e.g. en-US.; Default: “en-US”
SpeechToTextInput
Field Type Required Description and constraints
audio AsrAudioInput Yes Audio source to transcribe.
options AsrRequestOptions or null No Optional transcription feature toggles.; Default: null
Input schema, including the MCP wrapper
{
"$defs": {
"AsrAudioInput": {
"description": "Audio source — resolved to raw bytes for HTTP submit.",
"properties": {
"audio_data": {
"anyOf": [
{
"type": "string"
},
{
"type": "null"
}
],
"default": null,
"description": "Base64 audio bytes (max 200 MB decoded). Exactly one audio source."
},
"audio_file_path": {
"anyOf": [
{
"type": "string"
},
{
"type": "null"
}
],
"default": null,
"description": "Absolute local file path (stdio only, max 200 MB). Exactly one audio source."
},
"audio_format": {
"description": "Format of the source audio.",
"enum": [
"wav",
"mp3",
"ogg",
"raw",
"flac"
],
"type": "string"
},
"audio_url": {
"anyOf": [
{
"type": "string"
},
{
"type": "null"
}
],
"default": null,
"description": "HTTPS URL of the audio (max 200 MB). Exactly one audio source."
}
},
"required": [
"audio_format"
],
"type": "object"
},
"AsrRequestOptions": {
"description": "Optional transcription feature toggles.",
"properties": {
"enable_itn": {
"anyOf": [
{
"type": "boolean"
},
{
"type": "null"
}
],
"default": null,
"description": "Inverse text normalization (e.g. spoken numbers to digits)."
},
"enable_punc": {
"anyOf": [
{
"type": "boolean"
},
{
"type": "null"
}
],
"default": null,
"description": "Add punctuation to the text."
},
"language": {
"default": "en-US",
"description": "BCP-47 language code, e.g. en-US.",
"type": "string"
}
},
"type": "object"
},
"SpeechToTextInput": {
"description": "Input for the ``speech_to_text`` tool.",
"properties": {
"audio": {
"$ref": "#/$defs/AsrAudioInput",
"description": "Audio source to transcribe."
},
"options": {
"anyOf": [
{
"$ref": "#/$defs/AsrRequestOptions"
},
{
"type": "null"
}
],
"default": null,
"description": "Optional transcription feature toggles."
}
},
"required": [
"audio"
],
"type": "object"
}
},
"additionalProperties": false,
"properties": {
"input": {
"$ref": "#/$defs/SpeechToTextInput"
}
},
"required": [
"input"
],
"type": "object"
}
Output schema, including nested definitions
{
"$defs": {
"TranscriptionResult": {
"description": "Full transcription result returned by the speech_to_text tool.",
"properties": {
"duration_ms": {
"anyOf": [
{
"type": "integer"
},
{
"type": "null"
}
],
"default": null,
"description": "Total audio duration in milliseconds.",
"title": "Duration Ms"
},
"text": {
"description": "Full transcript text.",
"title": "Text",
"type": "string"
},
"utterances": {
"description": "Utterance-level segments with timestamps.",
"items": {
"$ref": "#/$defs/TranscriptionUtterance"
},
"title": "Utterances",
"type": "array"
}
},
"required": [
"text"
],
"title": "TranscriptionResult",
"type": "object"
},
"TranscriptionUtterance": {
"description": "An utterance segment with timing, words, and speaker metadata.",
"properties": {
"channel_id": {
"anyOf": [
{
"type": "string"
},
{
"type": "null"
}
],
"default": null,
"description": "Audio channel identifier if channel split is enabled.",
"title": "Channel Id"
},
"end_time_ms": {
"anyOf": [
{
"type": "integer"
},
{
"type": "null"
}
],
"default": null,
"description": "End time in milliseconds.",
"title": "End Time Ms"
},
"speaker_id": {
"anyOf": [
{
"type": "string"
},
{
"type": "null"
}
],
"default": null,
"description": "Speaker label if diarization is enabled.",
"title": "Speaker Id"
},
"start_time_ms": {
"anyOf": [
{
"type": "integer"
},
{
"type": "null"
}
],
"default": null,
"description": "Start time in milliseconds.",
"title": "Start Time Ms"
},
"text": {
"default": "",
"description": "Utterance text.",
"title": "Text",
"type": "string"
},
"words": {
"description": "Word-level detail within this utterance.",
"items": {
"$ref": "#/$defs/TranscriptionWord"
},
"title": "Words",
"type": "array"
}
},
"title": "TranscriptionUtterance",
"type": "object"
},
"TranscriptionWord": {
"description": "A single word with timing and confidence.",
"properties": {
"confidence": {
"anyOf": [
{
"type": "number"
},
{
"type": "null"
}
],
"default": null,
"description": "Recognition confidence (0.0-1.0).",
"title": "Confidence"
},
"end_time_ms": {
"anyOf": [
{
"type": "integer"
},
{
"type": "null"
}
],
"default": null,
"description": "End time in milliseconds.",
"title": "End Time Ms"
},
"start_time_ms": {
"anyOf": [
{
"type": "integer"
},
{
"type": "null"
}
],
"default": null,
"description": "Start time in milliseconds.",
"title": "Start Time Ms"
},
"text": {
"default": "",
"description": "Word text.",
"title": "Text",
"type": "string"
}
},
"title": "TranscriptionWord",
"type": "object"
}
},
"description": "Output for the ``speech_to_text`` tool.",
"properties": {
"log_id": {
"anyOf": [
{
"type": "string"
},
{
"type": "null"
}
],
"default": null,
"description": "Provider-side log ID for troubleshooting.",
"title": "Log Id"
},
"result": {
"$ref": "#/$defs/TranscriptionResult",
"description": "Full transcription result with text, utterances, and timing."
}
},
"required": [
"result"
],
"title": "SpeechToTextOutput",
"type": "object"
}
  • Missing tool: check provider credentials and feature flags, then restart the connection.
  • Invalid arguments: keep the correct wrapper and follow current model limits from discovery. A valid schema alone does not prove the chosen model supports the combination.
  • A job completes with an error: inspect the original tool result, not only the outer job status. Never automatically resubmit potentially billable work.
  • Local output writes from jobs require server-configured OUTPUT_ROOTS; HTTP cannot write to a client’s local disk.

See configuration, background execution, and artifact delivery.