D&D session transcription that already knows your characters’ names

How to transcribe a D&D session in three steps

Only the first step needs you at the table; the rest runs in the background and you can close the tab. The in-browser session recorder uses your device’s microphone, not the tab or system audio, so on an online game with headphones on it hears only you. For those games, record the call itself (how, on Discord) and upload that file instead.

1. Record the session in your browser, or upload the file you already have: a phone voice memo, a Craig mix or an OBS recording, a Zoom recording, even a video. One file can be up to 500 MB and 8 hours, and a file with several audio tracks has them mixed so every voice is transcribed.
2. AI transcription runs on OpenAI’s Whisper, with your campaign’s names handed to it first. You get an email when the transcript, the wiki entries and the recap are ready.
3. Read the timestamped transcript, fix anything it misheard, and look at what was built from it: new and updated wiki entries, a recap, and the session’s place on the campaign timeline.

Why AI transcription mangles fantasy names, and what fixes it

Speech recognition is a prediction problem. A model guesses the most probable word given the sound and everything before it, and "Thordak Ironforge" is not a probable word sequence in any general-purpose training corpus. You get "Thor Dak iron forge", or "thought Dak", and you get it differently on different sessions, so the same NPC ends up under four spellings and your notes stop being searchable.

The fix is not a better general model. It is telling the model what nouns to expect. DM Scribe builds a vocabulary prompt for every transcription from your campaign’s own data: the glossary you curate, your player characters and their aliases, the 200 most recently updated wiki entries, and the nicknames you have confirmed.

That prompt is budgeted in tokens, which is a real constraint rather than a marketing number — the transcription API honours roughly the final 224 tokens of a prompt, so the builder counts token cost instead of trusting a character count (with 850 characters as a hard ceiling on top). Recently updated entries win the space, on the assumption that what you touched last is what is currently on screen. If building the prompt fails for any reason, it falls back to empty rather than sending something malformed, and transcription proceeds normally.

When a name still comes out wrong, fix it once

A vocabulary prompt only nudges the model. A name nobody has typed yet — the NPC the party met an hour ago — still comes back spelled by sound, sometimes two ways in one session. Fix names handles that in one step: give it the wrong spelling and the right one, check the preview, and it corrects the recap, the transcript and the wiki, for that session or for the whole campaign.

Then it remembers the spelling. From that point every new transcript has the correction applied before it is saved, and so does every recap, so the same mistake does not come back next week. Remembered spellings also go into the vocabulary prompt, and the dialog suggests likely misspellings of names you have already confirmed, such as your player characters and your glossary, for you to check. It is on every plan, and only the GM can use it.

What you get back

One transcript per uploaded recording, stored against the session, with per-segment timestamps preserved from the transcription engine. Turn timestamps on to see when something was said; copy the whole thing to the clipboard, or download it as a text file you own.

• Timestamped segments, offset correctly across a split recording
• Readable in the app without downloading anything
• Copy to clipboard, or download as a plain .txt file
• Feeds the wiki, the recap, the timeline and Ask Your Campaign automatically

Long sessions are the normal case, not the edge case

A four-hour session is a large audio file and most transcription APIs will not take it in one request. DM Scribe splits on two levels so you never have to think about it. Sessions over four hours are divided into chunks that are transcribed separately and reassembled under one session; separately, any audio over roughly twenty minutes is split again to fit the transcription API’s per-request limit.

Both are invisible from your side. You upload one file and get one transcript, with timestamps offset correctly across the seams. Transcription runs as a background job in the cloud, so you can close the tab and walk away — you get an email when the transcript, the wiki entries and the recap are all ready.

Audio arrives in whatever your phone, recorder or virtual tabletop produced. It is converted to a normalised WAV before transcription, so you are not converting formats yourself.

If you already have the text, skip the audio entirely

A transcript from a Discord bot, another transcription tool or a play-by-post game can be imported as .txt or .md, or pasted in — the full chain still runs and it spends no transcription hours. See importing a transcript you already have.

Transcribing D&D sessions: short answers

Is there an app for transcribing D&D sessions? You do not need one. DM Scribe runs in the browser on desktop, tablet or phone, the recorder is part of the same web app, and anything you record with something else can be uploaded as a file.

Can AI transcribe a whole four-hour session? Yes — long sessions are the normal case, as the section above explains. Two honest limits: the transcript does not say who is speaking, and the audio sets the ceiling, because no model recovers speech the microphone never picked up clearly.

What about fantasy names? That is what the vocabulary prompt is for. Glossary terms go into it first, ahead of the wiki, and a term can carry a pronunciation hint — Wendelbrie (WEN-del-bree) — so the model has the spelling before it hears the word. When a name still comes out wrong, Fix names corrects it everywhere and remembers the spelling, so the fix carries into the next session.

Which plans include it

Included on every plan, including the free Adventurer tier, which gets 2.5 hours a month plus 3 one-time bonus hours at signup. Transcript import is on every plan too — 2 a month on Adventurer, 10 on Dungeon Master, unlimited on Guild Master.

Full pricing