TRANSCRIPT EDITING REVIEW WORKSHEET Checked October 10, 2026 | Burki free resource Guide: https://burki.dev/blog/elevenlabs-transcript-editing-original-record This is an unperformed review template. The resource is free; provider usage is not promised free. Burki currently has no native Scribe recognition selector or transcript-editing toggle. Keep secrets and unapproved customer data out of this worksheet. 1. SCOPE AND SOURCE Owner / reviewer: Permitted source record reference (not a public audio link): Retention / access requirements: Batch or realtime: Exact model / SDK version / processing date: Exact edit instruction and instruction version: Intended change (date / time / unit / abbreviation / other): Changes the edit must never make: 2. THREE SEPARATE RECORDS Original recognized text: Original word-time reference, when available: Edited proposal: Business field name and proposed value: Confirmation evidence / unresolved question: Completed action receipt, if any: Review decision: accepted formatting / rejected change / needs clarification / unavailable Reviewer / time / reason: 3. FICTIONAL WORKED CASE Original example: Could I collect it on the fourteenth of November twenty twenty-six? Sorry, the fifteenth. The return would be December the second. Starter instruction to evaluate: Normalize fully specified dates to YYYY-MM-DD. Preserve corrections, negations and uncertainty in the text. When a year is not stated or is ambiguous, leave that date in words. Do not add a reservation, approval or promise that the speaker did not make. Expected review criteria (not generated provider results): - Preserve the correction from November 14 to November 15. - Collection request may be recorded as 2026-11-15 only with its request/confirmation status. - Do not invent the return year; clarify it before requiring a complete date. - A requested collection is not a confirmed reservation. 4. APPLICATION FAILURE CASES Case | observed result | original retained? | edit state | business field blocked? | reviewer Ordinary complete date: Corrected date: Ambiguous year: Failed batch edit: Missing realtime edit event: Consecutive repeated identical source text: Edits arrive in reverse order: Realtime local simulation, not a provider test: Commit A = November fourteenth. Commit B = Sorry, November fifteenth. Edit B arrives before edit A. Later commit C has no edit event. Record whether association is unique. If ambiguous, keep it unresolved. Never attach the latest arriving edit blindly to the latest caller turn. Original timestamps must not be presented as alignment for rewritten text. 5. BUDGET AND PLAN Provider base transcription rate / unit / account plan / checked date: Editing rate / minimum duration rule / checked source: Batch requests and durations: Realtime committed segments and durations: Failed attempts / retries: Storage / application / other charges: Approved test budget (if separately authorized): Actual charge readback after your own authorized run: Current provider docs: editing premium is 30%; minimum 10 seconds per batch request, or 10 seconds per realtime committed transcript. Confirm current plan terms before use. Twenty two-second realtime commits have a 200-second editing minimum, not 40 seconds. This is illustrative arithmetic, not a bill or a base-transcription estimate. 6. BURKI REVIEW BOUNDARY Call reference: Available original transcript / post-call result: Captured field compared with conversation: Action receipt checked separately: Missing or incomplete evidence: No Scribe upload or editing configuration is performed by this worksheet. PRIMARY REFERENCES https://elevenlabs.io/docs/changelog https://elevenlabs.io/docs/eleven-api/guides/how-to/speech-to-text/batch/transcript-editing https://elevenlabs.io/docs/eleven-api/guides/how-to/speech-to-text/realtime/transcript-editing https://elevenlabs.io/docs/api-reference/speech-to-text/convert https://elevenlabs.io/pricing