Pre-Recorded Transcription#

Gladia's pre-recorded API transcribes audio and video files asynchronously.

SDK-first: always use the official SDK — see gladia-sdk-integration for policy, setup, and fallback criteria.

When to Use#

• Existing audio/video files or URLs (including social/video links)
• Batch or asynchronous transcription workflows
• Pre-recorded-only features: diarization, PII redaction, subtitles

When NOT to use: If the user needs real-time / live transcription of a stream, microphone, or ongoing audio feed, use the gladia-live-transcription skill instead. Live transcription uses WebSocket sessions, not the pre-recorded API. If the user has gladia-cli installed and wants a quick terminal workflow without writing code, use gladia-using-cli instead.

References#

Consult these resources as needed:

• ./references/transcription-options.md -- Full options (JS + Python)
• ./references/managing-jobs.md -- get, list, getFile, delete
• ./references/delivery-and-response.md -- Response shape and events
• ../gladia-audio-intelligence/SKILL.md -- Feature availability and config
• ../gladia-using-cli/SKILL.md -- Terminal workflow with gladia-cli
• ../gladia-sdk-integration/SKILL.md -- Setup, config, SDK vs raw API
• ../gladia-sdk-integration/references/sdk-versions.md -- Current SDK versions
• ../gladia-troubleshooting/SKILL.md -- Errors and diagnostics

API Endpoints (reference — prefer SDK methods instead)#

EndpointMethodSDK equivalent/v2/uploadPOSTtranscribe() auto-uploads local files/v2/pre-recordedPOSTcreate() / transcribe()``/v2/pre-recordedGETlist()``/v2/pre-recorded/:idGETget() / poll() / transcribe()``/v2/pre-recorded/:idDELETEdelete()``/v2/pre-recorded/:id/fileGETgetFile()

Workflow#

Recommended (SDK)#

The SDK transcribe() method handles upload, job creation, and polling in one call. Use this by default.

typescript

const result = await client.preRecorded().transcribe("./audio.mp3", {
  language_config: { languages: ["en"] },
  diarization: true,
});

console.log(result.result?.transcription?.full_transcript);

python

result = client.prerecorded().transcribe(
    "audio.mp3",
    {"language_config": {"languages": ["en"]}, "diarization": True},
)

print(result.result.transcription.full_transcript)

Audio input can be a local file path, HTTP(S) URL, social/video URL, or binary file object. For full input types, see gladia-sdk-integration.

Fallback (raw REST — only when SDK is not feasible)#

Use raw REST only when SDK use is not possible.

1. Upload (if local file): POST /v2/upload with multipart form data → get audio_url
2. Create job: POST /v2/pre-recorded with audio_url and config → get id
3. Poll: GET /v2/pre-recorded/:id until status: "done" (or use webhooks/callbacks)
4. Parse results: Extract transcription, diarization, translation, etc. from response

Managing Jobs#

Use SDK methods for post-processing operations:

• JavaScript: client.preRecorded().get(id), .list(filters), .getFile(id), .delete(id)
• Python: client.prerecorded().get(id), .list(filters), .get_file(id), .delete(id)

For full JS/Python examples, pagination filters, and REST equivalents, see ./references/managing-jobs.md.

Transcription Options#

All options are passed as the second argument to transcribe(). Key options:

OptionDescriptionlanguage_configExpected languages, code switchingdiarizationSpeaker identification (pre-recorded only)translationTranslate to target languagessummarizationGenerate bullet points or paragraph summarysubtitlesGenerate SRT/VTT filespii_redactionRedact PII (pre-recorded only)audio_to_llmRun custom LLM prompts on transcriptcallback_urlAsync webhook delivery

For full option details, see ./references/transcription-options.md. For audio intelligence config, see gladia-audio-intelligence. For client-level retry/timeouts, see gladia-sdk-integration.

Response and Delivery#

For full response JSON and event names, see ./references/delivery-and-response.md.

Limits and Specifications#

ConstraintValueMax file size1000 MBMax duration135 minutes (120 min for YouTube)Enterprise max duration4h15Concurrency (paid)25 concurrent jobsConcurrency (free)3 concurrent jobs

Polling Best Practices#

The SDK handles polling automatically — transcribe() polls until the job completes with configurable interval and timeout:

typescript

const result = await client.preRecorded().transcribe(audio, options, {
  interval: 5000, // Poll every 5s
  timeout: 600000, // Timeout after 10 minutes
});

If using raw REST instead of the SDK:

• Use webhooks or callbacks instead of polling when possible
• If polling, implement exponential backoff (start at 3s, max 30s)

Common Mistakes#

• Code switching without language list: enabling code_switching: true with empty languages triggers 100+ language evaluation. Always provide 3-5 expected languages.
• Polling without backoff: rapid polling wastes requests and may trigger 429s. The SDK handles this; for raw REST, use webhooks or exponential backoff.
• Expecting live-only features: diarization, PII redaction, and subtitles are pre-recorded only — not available in live mode.
• Wrong audio file path: the audio download endpoint is /v2/pre-recorded/:id/file, not /v2/pre-recorded/:id/audio.

For the full list of gotchas and diagnostics, see the gladia-troubleshooting skill.

Further Reading#

• Pre-recorded quickstart
• Audio intelligence overview
• API reference: init