Transcription¶

The Scribe's Art¶

TaleKeeper uses on-device speech recognition optimized for Apple Silicon to transcribe your recordings. Everything runs locally — your audio never leaves your machine.

Viewing the Transcript¶

Switch to the Chronicle tab (2) to see your full transcript.

Chronicle tab showing timestamped transcript segments with color-coded speaker names, audio player, and search bar

Each segment shows:

Timestamp — when the words were spoken
Speaker — who said it (assigned by diarization)
Text — what was said

Click to Seek

Click any transcript segment to jump to that moment in the audio player. As audio plays, the active segment is highlighted with a gold border and the transcript auto-scrolls to follow along. Navigation works both ways — click a segment to seek, or let playback drive the scroll.

Search and Filter¶

The search bar at the top of the Chronicle tab lets you filter transcript segments. It matches against both text content and speaker names — type a character name to see only their lines.

A match count shows how many segments match your query. Click Clear to reset.

Copying and Downloading¶

Copy a line: hover over any segment to reveal a clipboard icon. Clicking it copies the segment with its timestamp and speaker name — ready to paste into Discord, notes, or a blog.
Download transcript: click the download icon in the search bar to export the full transcript as a .txt file.

Color-Coded Speakers¶

Each speaker is assigned a unique color that's consistent throughout the transcript. This makes it easy to visually follow who's speaking, even in long sessions with many participants.

Smart Segment Splitting¶

When a single transcript segment contains two different speakers (common during rapid back-and-forth exchanges), TaleKeeper automatically splits it so each speaker gets their own line. This happens behind the scenes during speaker identification — you just see clean, correctly attributed segments.

Volume Normalization¶

Players sitting farther from the microphone can be harder to detect. TaleKeeper automatically adjusts for volume differences so that quiet speakers are identified just as reliably as loud ones.

Crosstalk Indicators¶

When multiple speakers talk over each other, those segments appear at reduced opacity with an italic [crosstalk] label. This makes it easy to spot moments of overlapping speech without cluttering the readable transcript.

Whisper Models¶

The model affects speed and accuracy. Configure it in Settings.

Model	Speed	Accuracy	Best For
`tiny`	~30 sec / 10 min audio	Lower	Quick previews, testing
`base`	~1 min / 10 min audio	Fair	Short sessions
`small`	~2 min / 10 min audio	Good	Most sessions
`medium`	~3 min / 10 min audio	Very Good	Balanced option
`distil-large-v3`	~2 min / 10 min audio	Excellent	Recommended default
`large-v3`	~5 min / 10 min audio	Best	Critical recordings, accented speech

Noise Filtering

Before transcribing, TaleKeeper automatically filters out silences, background music, and non-speech noise. This makes transcription faster and more accurate.

Long Sessions

For longer recordings, TaleKeeper automatically handles them in sections to ensure nothing is missed. You don't need to do anything — it's all handled for you.

Language Support¶

TaleKeeper supports 98 languages out of the box. Set the language at the campaign or session level, and transcription, summaries, and session names will all respect it.

Common languages: English, Spanish, French, German, Japanese, Korean, Chinese, Hebrew, Arabic, Portuguese, Italian, Russian, and many more.

Next: Re-run Transcription →