Verbit platform help — enterprise AI+human transcription, live captioning, and accessibility (ADA/WCAG/CVAA) for education, legal, media, and government. Use when setting up Verbit live captioning for Zoom or Teams events, uploading recorded audio for human-verified transcription, integrating the Verbit API (Insights, Live Booking, Post-Production, Search) into a transcript pipeline, comparing Verbit vs Rev vs 3Play Media vs Sonix for enterprise transcription, troubleshooting ASR punctuation or speaker diarization errors, choosing between self-service and enterprise plans, or configuring SmartPlayer for accessible video playback. Do NOT use for picking a sales meeting note-taker (use /sales-note-taker) or reviewing a call for coaching (use /sales-call-review).
Standardmäßig ist der Prompt ausgewählt, der zuerst die Quelle prüft. Sie können zu einem direkten Befehl wechseln oder eine lokale Kopie herunterladen.
Quelldateien prüfen
Lesen Sie SKILL.md und alle von SkillsMP angezeigten Begleitdateien, bevor Sie sich für eine Installation entscheiden.
Mit Codex oder Claude installieren Kopieren Sie diesen Prompt, fügen Sie ihn in Codex, Claude oder einen anderen Assistant ein und lassen Sie die Skill-Seite prüfen und installieren.
Ein direkter Befehl überspringt den Prüf-Prompt. Prüfen Sie die Quelle, bevor Sie ihn ausführen.
Verbit platform help — enterprise AI+human transcription, live captioning, and accessibility (ADA/WCAG/CVAA) for education, legal, media, and government. Use when setting up Verbit live captioning for Zoom or Teams events, uploading recorded audio for human-verified transcription, integrating the Verbit API (Insights, Live Booking, Post-Production, Search) into a transcript pipeline, comparing Verbit vs Rev vs 3Play Media vs Sonix for enterprise transcription, troubleshooting ASR punctuation or speaker diarization errors, choosing between self-service and enterprise plans, or configuring SmartPlayer for accessible video playback. Do NOT use for picking a sales meeting note-taker (use /sales-note-taker) or reviewing a call for coaching (use /sales-call-review).
Read references/platform-guide.md for the full platform reference — modules, pricing, integrations, data model, workflows.
If the question involves the API, also read references/verbit-api-reference.md.
Answer the user's question using only the relevant section. Don't dump the full reference.
Step 4 — Actionable guidance
You no longer need the platform guide — focus on the user's specific situation.
Plan selection: Self-Service ($29/mo) for occasional AI-only transcription up to 20 hrs/mo. Enterprise for human-verified accuracy, compliance (HIPAA BAA, SOC 2), SLA guarantees, and volume pricing. No mid-market tier exists — if you need more than 20 hrs/mo but don't need enterprise contracts, compare Sonix ($10/hr PAYG) or Rev ($0.25/min AI).
API selection: Use Post-Production API for batch upload workflows. Use Live Booking API for scheduled live captioning events. Use Insights API (Gen.V) for AI-generated summaries, keywords, and quizzes from completed transcripts — note this API lives on a separate host, https://ai.verbit.co/api/v2/insights, not the main api.verbit.co. For real-time streaming transcription over WebSocket, use the official Python SDK verbit-streaming-sdk (pip install --upgrade verbit-streaming-sdk). Base URLs: US api.verbit.co, EU eu.api.verbit.co, sandbox sandbox-api.verbit.co.
Accuracy optimization: Upload highest-quality audio. Request custom ASR model tuning on Enterprise. Use human review tier for legal/compliance content where 99%+ accuracy is required.
If you discover a gotcha, workaround, or tip not covered in references/learnings.md, append it there.
Gotchas
Best-effort from research — review these, especially items about plan-gated features and integration gotchas that may be outdated.
No mid-market pricing tier. Self-Service is $29/mo (20 hrs AI-only). Enterprise is custom (~$33K-$75K/yr). Nothing in between — mid-size teams are stranded.
Self-Service is AI-only transcription. No human review, no custom ASR models, no compliance certifications. Those require Enterprise.
ASR punctuation and grammar errors are common. Unnecessary spaces before punctuation, run-on sentences, missed capitalization — review and edit output before publishing.
Speaker diarization splits speakers incorrectly. Multiple speakers grouped under one label or one speaker split into several — worse with similar voices or cross-talk.
Bearer tokens expire after 24 hours. API integrations must handle token refresh. Post-Production API alternatively supports API key auth (simpler for server-to-server).
Gen.V Insights require completed transcripts. You can't generate summaries or keywords until transcription processing finishes.
Rate limits are not published. Design with conservative throttling and exponential backoff. Monitor for 429 responses.
VITAC (Verbit subsidiary) has separate quality issues. If you're evaluating broadcast captioning, VITAC reviews cite inconsistent quality and missed deliverables.
Related skills
/sales-note-taker — Comparing AI meeting note-takers and conversation intelligence platforms, or choosing batch transcription tools (Sonix, TranscribeMe, Rev)
/sales-sonix — Sonix platform help (batch AI transcription, translation, subtitles)
/sales-transcribeme — TranscribeMe platform help (human+AI hybrid transcription)
/sales-integration — General tool integration patterns (Zapier, webhooks, iPaaS)
/sales-do — Not sure which skill to use? The router matches any sales objective to the right skill. Install: npx skills add sales-skills/sales --skill sales-do
Examples
Example 1: Comparing Verbit vs Sonix for education transcription
User says: "I need to transcribe 500 hours of lecture recordings per semester for accessibility compliance. Should I use Verbit or Sonix?"
Skill does:
Notes the volume (500 hrs/semester) and compliance requirement (ADA/WCAG)
Compares: Sonix is $10/hr PAYG ($5K total) with AI-only output. Verbit Enterprise offers human-verified transcription with ADA/WCAG compliance certifications and LMS integrations (Blackboard, Canvas).
Recommends Verbit Enterprise if accuracy and compliance documentation matter. Sonix if budget is primary and human review can be done in-house.
Notes Verbit's Captivate ASR can be custom-tuned for academic terminology on Enterprise.
Result: Clear comparison based on volume, compliance needs, and budget.
Example 2: Setting up live captioning for Zoom events
User says: "I need live captions for our weekly town hall on Zoom — about 200 attendees."
Skill does:
Reads the platform guide for Live Booking API and Zoom integration details
Explains Verbit integrates natively with Zoom for live captioning — no separate bot needed
Walks through booking flow: create order via Live Booking API or dashboard, specify Zoom meeting ID, select captioning profile
Notes human captioners are available for Enterprise plans (higher accuracy) vs AI-only for Self-Service
Result: Step-by-step setup for Zoom live captioning with plan-appropriate options.
Example 3: Building a post-production transcript pipeline
User says: "I want to automatically transcribe every recording uploaded to our S3 bucket using the Verbit API."
Skill does:
Reads the API reference for Post-Production endpoints
Sketches the pipeline: S3 event → Lambda → POST /job/new + POST /job/add_media (S3 URL) + POST /job/perform_transcription → poll GET /job/info for status → GET /job/get_caption for completed transcript
Notes API key auth is simpler for server-to-server (vs OAuth Bearer tokens with 24hr expiry)
Flags: no webhook for job completion documented — must poll status endpoint
Result: Working pipeline architecture with Verbit-specific API flow.
Troubleshooting
ASR output has punctuation and grammar errors
Symptom: Transcripts have extra spaces before periods, missing capitalization, run-on sentences
Cause: Verbit's Captivate ASR engine produces raw output that needs post-processing — common across all ASR engines but more noticeable in Verbit's output per user reports
Solution: Use the Verbit editor to clean up output. On Enterprise, request custom ASR model tuning for your domain vocabulary. For critical content (legal, compliance), use human review tier instead of AI-only.
Speaker labels are wrong or missing
Symptom: Two speakers labeled as one, or one speaker split into 5+ labels
Cause: Speaker diarization struggles with similar voices, cross-talk, poor audio quality, and single-channel (mono) recordings
Solution: Upload stereo/multi-channel audio when possible. Review and merge/rename speakers in the editor. On Enterprise, request custom speaker model training. For critical content, use human review.
Post-Production API job stuck in processing
Symptom: Job status stays "processing" for hours after upload
Cause: Large files, queue backlog, or unsupported format. Human review tier has longer turnaround (hours to days, not minutes).
Solution: Check file format is supported (MP3, WAV, MP4, M4A, FLAC, OGG). For large files, use URL-based upload instead of direct upload. Check Verbit status page for service issues. On Enterprise, contact support for priority processing. AI-only jobs should complete in minutes — if not, the file may have processing errors.
