How to Transcribe Discord Voice Calls
Discord voice calls are widely used for gaming communities, study groups, project teams, and remote collaboration — but Discord does not offer built-in recording or transcription. To transcribe a Discord call, you need to record it with a third-party tool like Craig Bot or OBS, then upload the audio file to an AI transcription service. BrassTranscripts processes Discord call recordings in 1-3 minutes per hour of audio, with automatic speaker identification and 4 output formats (TXT, SRT, VTT, JSON).
This guide covers the two main methods for recording Discord calls, the audio formats each tool produces, and how to upload recordings to BrassTranscripts for transcription.
Quick Navigation
- Why Transcribe Discord Calls
- Method 1: Recording with Craig Bot
- Method 2: Recording with OBS
- Discord Built-In Recording: What It Can and Cannot Do
- Supported Audio Formats from Discord Recording Tools
- How to Upload Discord Recordings to BrassTranscripts
- Community Meeting Use Cases
- Tips for Better Discord Audio Quality
- Frequently Asked Questions
Why Transcribe Discord Calls
BrassTranscripts converts Discord voice call recordings into searchable, timestamped text with speaker labels — turning ephemeral voice conversations into permanent, shareable documentation that community leaders, study groups, and project teams can reference later.
Discord voice channels are increasingly used for serious collaboration beyond gaming. Community managers run weekly town halls. Study groups hold exam prep sessions. Open source project teams conduct sprint planning. In all these cases, the conversation disappears the moment the call ends unless someone records and transcribes it.
Common reasons to transcribe Discord calls:
- Meeting minutes for community governance and decision tracking
- Study notes from group tutoring or exam prep sessions
- Project documentation from sprint planning and design discussions
- Content creation — podcasters and interviewers who record on Discord
- Accessibility — providing text versions for deaf or hard-of-hearing community members
- Accountability — written records of agreements, votes, and action items
Manual transcription takes 4-6 hours per hour of audio according to industry standards documented by Rev.com. AI transcription reduces that to minutes. For a complete introduction to the process, see the getting started with AI transcription guide.
Method 1: Recording with Craig Bot
BrassTranscripts works well with Craig Bot recordings because Craig Bot produces high-quality FLAC or OGG audio files with separate tracks per speaker — ideal input for AI transcription with speaker identification.
Craig Bot is a purpose-built Discord bot that records voice channel audio. It is one of the most widely used recording bots on Discord, and its key advantage for transcription is multi-track recording: each participant gets their own separate audio track.
How to Set Up Craig Bot
- Invite Craig Bot to your Discord server from craig.chat
- Grant permissions — Craig Bot needs permission to join and listen in voice channels
- Join a voice channel and type
/joinin a text channel (or use the slash command in the voice channel) - Craig Bot joins and begins recording all participants
- Stop recording by typing
/stopor when everyone leaves the channel - Download recordings — Craig Bot sends a private DM with download links
Craig Bot Output Formats
Craig Bot offers several download options:
- FLAC (lossless) — Best quality for transcription. Produces a single mixed file or separate per-speaker files
- OGG (Vorbis) — Compressed but good quality. Smaller file sizes than FLAC
- Multi-track — Separate audio files for each speaker, useful for editing before transcription
- Mixed single track — All speakers combined into one file, ready to upload directly
For transcription, download the single mixed FLAC or OGG file. This gives BrassTranscripts one file with all speakers, and the automatic speaker identification will label each person in the output.
Important Notes About Craig Bot
- Craig Bot records server voice channels, not direct (1-on-1) DM calls
- Recordings are available for download for a limited time (typically 6 hours for the standard tier)
- The bot announces when it starts recording, so all participants are aware
- Craig Bot is operated by a third party — review their privacy policy for your use case
Method 2: Recording with OBS
BrassTranscripts accepts the MP4 and MKV files that OBS produces, making OBS a reliable recording option when Craig Bot is unavailable or when you need to record DM calls that Craig Bot cannot access.
OBS (Open Broadcaster Software) is an open-source tool primarily used for streaming and screen recording. It can capture Discord audio by recording your computer's system audio output.
How to Record Discord Audio with OBS
- Download and install OBS from obsproject.com
- Create a new Scene in OBS
- Add an Audio Output Capture source — this captures your system audio, which includes Discord
- Optional: Add an Audio Input Capture source — this captures your microphone so your own voice is also recorded
- Set output format — Go to Settings > Output > Recording and choose MKV or MP4
- Set audio quality — Under Settings > Audio, set the sample rate to 48kHz (Discord's native rate)
- Start your Discord call, then click "Start Recording" in OBS
- Stop recording when the call ends
- If you recorded in MKV — Use OBS's built-in "Remux Recordings" (File menu) to convert to MP4 for easier handling
OBS Audio Settings for Best Transcription Quality
- Sample rate: 48kHz (matches Discord's audio pipeline)
- Audio bitrate: 160kbps or higher
- Channels: Stereo
- Disable video if you only need audio — this creates much smaller files. In OBS, you can remove all video sources and keep only audio sources
OBS Limitations for Discord Recording
- Single mixed track — OBS captures all audio as one stream, unlike Craig Bot's per-speaker tracks. BrassTranscripts speaker identification still works, but separate tracks produce cleaner results
- Captures all system audio — Unless configured carefully, OBS will also capture notification sounds, music, and other application audio. Close other audio-producing applications during recording
- No automatic notification — Unlike Craig Bot, OBS does not announce to other participants that recording is in progress. Inform participants manually and obtain consent where required
For detailed guidance on recording settings that improve transcription results, see the audio quality guide.
Discord Built-In Recording: What It Can and Cannot Do
Discord does not provide built-in call recording or transcription for regular users — the only native capture feature is Clips, which records the last 30 seconds of a voice channel for short highlights, not full conversations.
As of March 2026, Discord's recording capabilities are limited:
- Clips (introduced 2023) — Captures the last 5-30 seconds of a voice channel. Designed for gaming highlights, not meeting documentation. Clips are short video recordings, not audio transcription
- No full call recording — Discord has no "Record" button for voice channels or DM calls
- No built-in transcription — Unlike Zoom or Microsoft Teams, Discord does not offer real-time or post-call transcription
- Live captions — Discord added experimental live captions for accessibility, but these are real-time only and do not produce a downloadable transcript
Bottom line: You need a third-party tool (Craig Bot, OBS, or similar) to record Discord calls for transcription. For a comparison of how other platforms handle built-in transcription, see the meeting transcription guide.
Supported Audio Formats from Discord Recording Tools
BrassTranscripts supports 11 audio and video formats, covering every output format produced by Craig Bot and OBS — upload directly without conversion.
Here is what each tool produces and whether BrassTranscripts accepts it:
| Tool | Output Format | BrassTranscripts Compatible |
|---|---|---|
| Craig Bot | FLAC | Yes |
| Craig Bot | OGG | Yes |
| OBS | MP4 | Yes |
| OBS | MKV | Yes |
| OBS | FLV | Yes |
| OBS (remuxed) | MP4 | Yes |
File size limit: 250MB maximum per file. Most Discord call recordings fall well under this — a 2-hour FLAC recording from Craig Bot typically ranges from 50-150MB depending on the number of speakers and audio complexity.
Duration limit: 2 hours maximum per file. For longer calls, split the recording or use multiple upload batches.
For the complete list of supported formats including WAV, MP3, M4A, WEBM, and others, see the file formats guide.
How to Upload Discord Recordings to BrassTranscripts
BrassTranscripts processes Discord call recordings through a straightforward upload workflow — upload the file, review a 30-word preview, pay per file, and download the completed transcript with speaker labels in minutes.
Step-by-Step Upload Process
- Go to brasstranscripts.com and create an account (takes about 30 seconds)
- Upload your recording — Drag and drop or click to select your FLAC, OGG, MP4, or MKV file
- Processing begins — BrassTranscripts processes the audio in 1-3 minutes per hour of audio
- Review the 30-word preview — A preview of the transcript is shown before payment so you can verify quality
- Pay per file — $2.50 for recordings up to 15 minutes, $6.00 for recordings between 16 and 120 minutes
- Download your transcript — Choose from 4 output formats: TXT (plain text), SRT (subtitles), VTT (web captions), or JSON (structured data)
Speaker Identification in Discord Transcripts
BrassTranscripts automatically identifies different speakers in the recording and labels them as Speaker 0, Speaker 1, Speaker 2, and so on. After downloading, you can replace these generic labels with actual names using find-and-replace in any text editor.
For a deeper look at how speaker identification works and tips for improving accuracy, see the speaker identification guide.
Tip for Craig Bot users: If Craig Bot gives you separate per-speaker files, you can either upload the single mixed track (speaker identification will separate voices automatically) or upload individual speaker files separately if you only need specific participants transcribed.
Community Meeting Use Cases
BrassTranscripts serves Discord communities that need permanent records of voice conversations — from gaming guild governance to academic study groups to open source project coordination.
Discord voice calls serve a wide range of purposes beyond casual conversation. Here are common scenarios where transcription adds value:
Gaming Communities and Guild Management
- Officer meetings — Document decisions about guild policies, raid schedules, and member disputes
- Town halls — Create searchable records of community announcements and Q&A sessions
- Recruitment interviews — Keep records of prospective member interviews for officer review
- Event planning — Capture logistics discussions for tournaments and community events
Study Groups and Academic Collaboration
- Exam prep sessions — Create study notes from group review sessions that all members can reference
- Tutoring sessions — Generate written records that students can revisit after the session
- Research discussions — Document brainstorming and literature review conversations
- Group project coordination — Track task assignments and deadlines discussed in voice
Project Teams and Open Source
- Sprint planning — Record and transcribe planning meetings for distributed teams
- Design reviews — Create written records of feedback and decisions on designs or code
- Onboarding sessions — Transcribe orientation calls for new team members to reference later
- Retrospectives — Document what went well and what needs improvement
Content Creators and Podcasters
- Interview recordings — Many podcasters conduct interviews on Discord due to its audio quality and ease of use
- Collaboration calls — Document creative discussions between co-hosts or collaborators
- Show notes — Generate written content from podcast recordings for blog posts and social media
Tips for Better Discord Audio Quality
BrassTranscripts produces the best transcription results when the input audio is clean — following these Discord-specific audio practices directly impacts transcript accuracy and speaker identification quality.
Discord compresses audio during transmission, so the recording quality depends heavily on each participant's setup and Discord's audio settings.
Discord Audio Settings
- Enable Noise Suppression — Discord's built-in noise suppression (Settings > Voice & Video) removes background noise. Use the "Standard" setting for the best balance between noise removal and voice clarity
- Use Push-to-Talk or Voice Activity with a proper threshold — This prevents background noise from triggering transmission when a participant is not speaking
- Set Input Sensitivity correctly — If using Voice Activity, adjust the sensitivity slider so it does not pick up keyboard clicks, fan noise, or room ambiance
- Disable Automatic Gain Control if participants have consistent microphone levels — AGC can create volume inconsistencies that affect transcription
Participant Best Practices
- Use headphones — This eliminates echo from speakers playing other participants' voices back into the microphone
- Use a dedicated microphone — A USB microphone or headset microphone produces clearer audio than a laptop's built-in microphone
- Minimize background noise — Close windows, mute when not speaking, and avoid typing on a mechanical keyboard near the microphone
- Maintain consistent distance from the microphone — Moving toward and away from the mic creates volume fluctuations
Recording-Specific Tips
- Close other audio applications before recording with OBS to prevent notification sounds and music from contaminating the recording
- Test your recording setup before an important call — do a short test recording and verify it captures all participants clearly
- Monitor audio levels during recording if using OBS — watch the audio meters to ensure levels stay in the green/yellow range without clipping into red
- Record in the highest quality available — Use FLAC from Craig Bot rather than OGG when file size is not a concern
For comprehensive audio optimization advice beyond Discord, see the audio quality guide.