assemblyai-transcription
Use this skill when an agent needs AssemblyAI for speech-to-text or speech-understanding work in media production, including pre-recorded, synchronous short-file, and real-time streaming transcription; speaker diarization or speaker identification; captions and subtitles; timestamps; language detection, code-switching, or transcript translation; audio intelligence such as summaries, chapters, topics, entities, key phrases, sentiment, content moderation, profanity filtering, and PII redaction; webhooks, scaling, rate limits, retention, security, consent, and QA for podcasts, interviews, caption
Profile →npx skills add calesthio/generative-media-skills --skill assemblyai-transcription --agent claude-code
Same command for any agent — swap --agent for codex, cursor, copilot.
Weekly change comes from our own snapshots, not the repository page — it measures attention, not adoption.
What it does
This skill enables transcription and speech-understanding tasks using AssemblyAI for media production. It covers pre-recorded and streaming transcription, diarization, captions/subtitles, timestamps, language detection, code-switching, translation, and various speech-understanding features (summaries, chapters, topics, entities, key phrases, sentiment, content moderation, profanity filtering, PII redaction). It also accounts for webhooks, scaling, rate limits, retention, security, consent, and QA for podcasts and related workflows. It prescribes choosing among surfaces (pre-recorded STT, sync STT, streaming STT) based on production needs, and it details model selections (Universal-3.5 Pro, Universal-2, streaming variants) with cost considerations. It includes inputs handling (audio upload, audio_url, start/end constraints), speaker separation/identity, captions/subtitles formats (SRT/VTT), language detection, translation, and guardrails like PII redaction and content moderation. It emphasizes production workflows: publish-ready artifacts, QA human-review for speaker labeling, and maintaining a job ledger. It also provides guidance on handling multi-channel vs. diarized outputs, and on exporting edit data (durable transcript JSON and EDL-like spans).
How it works
- Choose the product surface: pre-recorded STT (batch/async), sync STT (short clips), or real-time streaming STT over WebSocket. Use the appropriate endpoint flows: upload media to /v2/upload then submit audio_url, poll GET /v2/transcript/{id}, or leverage webhooks. For streaming, connect to wss://streaming.assemblyai.com/v3/ws and handle Begin, Turn, and Termination events with millisecond timings.
- Model and feature selection: pick Universal-3.5 Pro for high accuracy in English/other languages, or Universal-2 for broader language coverage or lower cost. For streaming, use Universal-3.5 Pro Streaming or other streaming variants. If Medical Mode is needed, set domain to medical-v1. Apply keyterms for domain terms to improve accuracy.
- Inputs and artifacts: upload local files to /v2/upload and pass the returned upload_url as audio_url; you may constrain transcripts with audio_start_from and audio_end_at. Retrieve word-level timestamps in the words array; use SRT/VTT endpoints for subtitles. Preserve transcript IDs and artifacts for auditing.
- Speaker handling: enable speaker_labels on pre-recorded transcripts for diarization; use speaker_options and mapping to names if available. For real-time, diarization and multichannel workflows are supported but with live constraints.
- Capabilities: enable translation, summarization, entity detection, sentiment, topic detection, key phrases, PII redaction, and content moderation as part of Speech Understanding, with appropriate guards and QA steps.
When to use it
- Use pre-recorded STT for finished media assets to maximize full-context transcripts for captions, chapters, and quote extraction.
- Use real-time STT during live events for captions or live agent assistance, then re-transcribe asynchronously for publication-grade transcripts.
- Use Sync STT only for short clips requiring immediate results with minimal lifecycle overhead.
What it can touch
- Tools: claude-code, codex, copilot, cursor are listed as declared tools.
- Endpoints and artifacts: /v2/upload, /v2/transcript/{id}, GET /v2/transcript/{transcript_id}/{subtitle_format}, streaming WebSocket at wss://streaming.assemblyai.com/v3/ws. audio_url, audio_start_from, audio_end_at parameters, transcript_id, subtitle outputs, and durable transcript JSON with reviewer notes.
Caveats
- Model names, pricing, add-ons, rate limits, and data-retention settings are volatile; re-check live docs before production use.
- PII redaction outputs redacted audio URLs are time-limited (24 hours) and redacted-audio creation is limited to original files smaller than 1 GB.
- Diarization quality depends on speech duration per speaker and audio quality; cross-talk and noise can reduce accuracy. Multichannel billing is per-channel.
- Real-time translation via streaming employs LLM Gateway rather than native streaming transcript fields.
- Requires human speaker-label QA before publishing captions or legal/medical notes.
# AssemblyAI transcription for media-production agents Use AssemblyAI when the production job needs reliable transcript artifacts from recorded or live speech, especially when the transcript must drive captions, clip selection, speaker turns, show notes, chaptering, translation, moderation, or compliance review. Treat this as a provider skill, not a generic ASR primer. Check the live AssemblyAI do
What does the assemblyai-transcription skill do?
Use this skill when an agent needs AssemblyAI for speech-to-text or speech-understanding work in media production, including pre-recorded, synchronous short-file, and real-time streaming transcription; speaker diarization or speaker identification; captions and subtitles; timestamps; language detection, code-switching, or transcript translation; audio intelligence such as summaries, chapters, topics, entities, key phrases, sentiment, content moderation, profanity filtering, and PII redaction; webhooks, scaling, rate limits, retention, security, consent, and QA for podcasts, interviews, caption
How do I install it?
Run `npx skills add calesthio/generative-media-skills --skill assemblyai-transcription --agent claude-code` — it drops the skill into your project so the agent can pick it up. Swap the --agent value for codex, cursor or copilot if you use one of those.
Where does this skill come from?
From calesthio/generative-media-skills, a repository with 112 stars. We read it straight from the repository tree rather than a submitted listing, so what you see here is what is actually published.
Is a popular skill a good skill?
Not necessarily. Stars measure attention, not adoption — a repository can trend for a week and be abandoned. That is why we show the weekly change from our own snapshots next to the total, instead of a single flattering number.