The two inputs
On the Session tab, next to the campaign picker:
- Speakers (the speaker icon) is the sound your PC plays: Discord voices, a video call, game audio. Pick the device you listen on, such as your headset.
- Mic (the microphone icon) is your own voice. Choose your microphone if you play on Discord, because Discord doesn’t play your own voice back to you. Choose None (already mixed in) if your voice is already part of what the speakers capture.
Playing in person? Set Mic to the microphone in the middle of the table and Speakers to None. One good microphone picks up the whole table. In-person transcripts don’t have speaker names.
Read the meters
Each input has a meter beside it while no session is running. Talk, or play some audio:
- hearing ✓ in green means it’s working.
- no signal means that input hears nothing. It’s the wrong device, or the sound is going somewhere else.
- can’t open means Windows wouldn’t give LoreMonkey that device. Pick another, or close the app that’s holding it.
Run Test setup
Press Test setup, then Start test, and say something out loud or play some game audio. It runs three checks:
- Audio: is your selected input being heard. “Very quiet” here means the audio isn’t reaching that device.
- Transcription: can it turn speech into text. It shows what it heard.
- AI: does LoreMonkey’s AI answer. This sends one tiny request.
Run it again with Run again after you change headphones, microphones or Discord’s output device.
Language and transcription quality
Above the live transcript:
- Language is the language spoken at your table. LoreMonkey transcribes and writes in it.
- Transcription quality is best left on Auto (recommended), which picks the most accurate model your PC can run in real time. Faster and Fastest help an older PC keep up, at some cost to accuracy.
Keeping the audio
Tick Record this session to keep the audio of your sessions. LoreMonkey shows how much space an hour takes. Recordings stay on your PC, cost nothing, and let you clean up a transcript afterwards with a more accurate model. You set the quality and how much space recordings may use in Settings → Recordings.