Soundcheck and recorded transcripts
Run consent-noticed recorded sessions and review supported recordings or transcripts within their verified visibility and persistence limits.
- Audience
- moderator · senior reviewer · team lead · server admin
- Availability
- premium preview · private beta access
- Maturity
- alpha · partial
Last reviewed
On this page
What Soundcheck records
Soundcheck creates a consent-noticed voice or Stage session with staff controls. Supported sessions can produce recordings and transcripts that authorized staff review in the dashboard or preserve as case evidence; this guide does not claim unsupported general live-transcript behavior.
Confirm purpose, access, and consent
Use the registered Soundcheck commands in Discord to configure, stage, start, inspect, and stop a supported session. Use Recorded Transcripts in the selected-server dashboard for later review when a stored session exists and is visible.
- A server with Soundcheck access and configured staff/consent surfaces.
- A clear purpose and participant notice appropriate to the supported session mode.
- Consent state, staff authority, and recording/transcription dependencies verified before capture.
Create, record, stop, and preserve
Create the consent-noticed session
Select the supported general or Stage mode and publish the implemented participant notice.
Check: Participants receive the consent controls and session scope before recording.
Confirm consent and access
Verify that a participant is in the source voice/Stage context and has granted consent before admitting or approving them to speak.
Check: Declined or missing consent blocks the protected participation path.
Record, stop, and check status
Authorized staff start and stop the supported session while monitoring visible session status.
Check: The record shows whether recording or transcription dependencies were missing.
Review and preserve
Open the stored transcript/recording through the server dashboard, respect visibility settings, and save only permitted material as evidence.
Check: Authorized reviewers can inspect a persisted artifact without widening its audience.
Soundcheck command examples
A supported session records consent and staff control state, then exposes only the recordings or transcripts that were actually produced and remain visible to the current server-scoped viewer.
/soundcheck status/soundcheck start reason:Witness interview case:104/soundcheck staff reason:Case #104 review meeting/soundcheck stage reason:Community town hall debate_queue:true debate_mode:Review/soundcheck stop session:104| Moment | Expected behavior |
|---|---|
| Session created | Discord shows the session scope, participant notice, staff controls, and session identifier. |
| Participant joins | The implemented consent controls explain that consent permits participation; it does not itself begin recording. |
| Recording starts | The session status indicates recording state and any transcription dependency that is unavailable. |
| Session stops | The final status distinguishes recording, transcript, and publication outcomes instead of promising every artifact. |
- Run a Stage Soundcheck where potential speakers consent before staff approves speaking.
- Open a recorded session in the dashboard and download a recording only if the route and viewer permission allow it.
- Save a transcript as case evidence while retaining its source identity and visibility limit.
Soundcheck modes and command responsibilities
| Command | Purpose | Verify |
|---|---|---|
| /soundcheck status | List or inspect active sessions before creating or stopping work. | Server, session ID, mode, owner, participants, consent, and recording state. |
| /soundcheck config | Change supported server-level Soundcheck enablement and limits. | Saved values and one short test session; do not infer general live transcription. |
| /soundcheck start | Create a supported general or case-linked consent-noticed session. | Reason, case reference, channel/workspace, notices, and session ID. |
| /soundcheck staff | Create or manage a staff-focused recorded review session. | Audience remains staff-only and participant expectations are clear. |
| /soundcheck stage | Run an in-place Stage Channel session with supported speaker/queue controls. | Stage identity, debate queue/mode, notices, speaker decisions, and recording state. |
| /soundcheck stop | Stop the intended active session and begin supported finalization. | Correct session ID and separate recording, transcript, publication, and storage outcomes. |
Consent, participation, and recording are separate states
- STAGE 01
Notice
Participants receive the implemented explanation of the session and consent controls.
- STAGE 02
Consent choice
A participant accepts or declines the supported participation/recording terms.
- STAGE 03
Admission or speaking
Staff use the current mode's controls to admit, approve, or manage participation.
- STAGE 04
Recording
An authorized staff action and available dependency start capture; consent alone does not start it.
- STAGE 05
Transcription
A separate provider/processing step may produce text; recording success does not guarantee transcript success.
- STAGE 06
Publication and review
Produced artifacts become available only through their supported visibility and storage routes.
| Question | Required answer before proceeding |
|---|---|
| What is the session for? | A specific moderation, interview, hearing, staff review, or Stage purpose supported by server policy. |
| Who can participate? | The current mode's staff/participant audience and speaker controls are understood. |
| Who has consented? | The implemented status shows consent for each protected participant; presence alone is not consent. |
| What is being captured? | Recording and transcription states are read separately. |
| Who can review the result? | Transcript/recording visibility and case evidence audience are known. |
Run and monitor a session
Prepare the environment
Confirm server feature access, staff permission, voice/Stage channel, participant notice, case link where relevant, and configured limits.
Check: The session has a clear owner and purpose.
Create one session
Use the matching start, staff, or stage command and keep the returned session identifier.
Check: Discord shows the intended session and staff controls.
Verify participant state
Confirm source-channel presence, consent choice, admission/speaker state, and any queue decision before protected participation.
Check: Declined or missing consent is not bypassed.
Monitor status
Check recording/transcription state, elapsed time, participant changes, and any visible dependency warning.
Check: Staff respond to a real status rather than assuming capture.
Stop and confirm
Stop the known session once and wait for finalization feedback.
Check: The record distinguishes stop success from later artifact processing.
Review recordings and transcripts
| Status | Meaning | Staff action |
|---|---|---|
| Recording available | Audio was captured and persisted through the supported storage path. | Use guarded playback/download and confirm viewer permission. |
| Recording missing/failed | No usable recording artifact was produced or persisted. | Document the failure; do not imply audio exists because the session ran. |
| Transcript processing | Text generation has not reached a final state. | Wait for the stored status; do not repeatedly stop or restart the completed session. |
| Transcript available | A stored transcript exists for the visible session. | Review speakers/timing and context, then preserve only necessary portions through evidence tooling. |
| Transcript partial/failed | Some or all transcription could not be produced. | State the gap and review the available recording or other evidence if permitted. |
| Hidden/denied | The current viewer lacks session or artifact access. | Hand off to an authorized reviewer; a copied link does not grant access. |
Voice Evidence in the dashboard is the review surface for stored sessions and artifacts. A session list row is not proof that every recording or transcript exists. Open the detail, read source/mode/status, and use the case/evidence link for consequential decisions.
Privacy limits and reserved voice features
- Keep recordings and transcripts inside guarded server, session, and Case surfaces.
- Do not repost full audio or transcript text into a wider staff channel to work around permissions.
- Collect only the session needed for the moderation purpose and follow the server's approved retention process.
- Treat local file or source unavailability as a documented gap, not proof that the session never occurred.
- Unreleased analysis, automatic transcripts, and general live-transcript behavior are not documented operating workflows.
A session ended without the expected artifact
What you see: The Discord session is closed, but recording, transcript, or playback is missing or failed.
What may be happening: Consent, recording dependencies, transcription provider, finalization, storage persistence, visibility, or a stale page can independently affect the artifact.
What to do
- Open the exact session detail and read each status separately.
- Confirm consent and source mode.
- Refresh after the supported processing period without recreating the session.
- Record the missing-artifact result and escalate configuration issues without attaching private audio/text.
Consent and artifact limits
Recording, transcription, live-caption settings, artifact storage, and dashboard visibility are distinct. A session can complete without every artifact. Local storage defaults do not promise indefinite or globally durable retention, and this guide does not advertise general live transcript behavior beyond verified session controls.
Recover a missing session artifact
The result is not what you expected
What you see: Recording or transcript output is missing, the session cannot start, or a stored session is hidden.
What may be happening: Consent, voice access, feature access, recording/transcription dependencies, visibility settings, or artifact persistence may block the expected result.
What to do
- Read the current Soundcheck status and consent state.
- Do not bypass a declined or missing consent state.
- Confirm server-scoped transcript visibility and viewer permission.
- Record the missing-artifact outcome and escalate configuration issues to a server administrator.
Share the session record or case evidence reference with authorized reviewers. Never repost recordings or transcripts into a broader channel simply to work around visibility.
Captions, recordings, and transcripts are separate
Live captions, recorded audio, generated transcript text, session metadata, and a Case-linked artifact are separate records. Starting a session does not promise that every artifact will be created, permanently retained, or automatically attached to a Case.
- Premium currently includes a monthly Soundcheck minute allowance; reservation and final usage can differ.
- Consent and participant notice remain required even when the UI shows a recording or caption indicator.
- Bearer links and exports must be treated as confidential copies, and provider-hosted media can expire.
- Case links and release behavior should be verified in the deployed version before relying on automatic attachment.