Archivist AI vs DM Scribe: the difference is what gets metered

The numbers, with sources

How we sourced this: Archivist’s figures were re-read from its own pricing page on 2026-09-07 and its input methods from its homepage, both linked per row. Annual pricing is not quoted because the pricing page does not mention an annual option — that is an absence we observed, not a claim that none exists. Storage limits and per-plan feature differences are likewise omitted: the page states "all features" on every tier without detailing them, so anything more specific would be our inference rather than their statement. This is a small, actively developed product and its pricing is the part most likely to have moved since the date above.

Sessions versus hours is the whole decision

Archivist’s plans are counted in sessions per month: 4 on Casual, 10 on Seasoned, 20 on Pro, 40 on Legend. A session is a session whether it ran for ninety minutes or six hours. Ours are counted in hours of audio: 2.5 a month free, 10 on the $7.99 plan, 30 on the $14.99 plan.

Work out your own number before reading any further. A weekly group playing four-hour sessions uses 4 sessions and about 16 hours in most months, 5 and 20 in the occasional five-week one. On Archivist that is the $10 Casual plan, plus a $2 extra session in the long months. On ours, 16 hours does not fit the $7.99 plan and you would be on the $14.99 one. For that table, Archivist is cheaper.

Now change one variable. A group that plays twice a week for two hours also uses 16 hours, but 8 sessions — which on Archivist needs the $20 Seasoned plan, and on ours is still $14.99. Same audio, opposite answer. Neither pricing model is sharper than the other; they are optimised for different tables, and you already know which one you are. (A third meter exists in this category too: Tabletop Arc counts minutes of audio, which is the same decision at a finer grain; GM Assistant meters hours too, sold as credits.)

What Archivist does that we do not

Named plainly, because a comparison page that finds no advantages in the other tool is not worth reading.

• Play-by-post and raw text ingestion. If your campaign partly lives in a Discord text channel, Archivist takes that directly. We accept a pasted or uploaded transcript on every plan — two a month free, ten on Dungeon Master, unlimited on Guild Master, none of them spending transcription hours — but a text-first campaign is still closer to what they are built around.
• Foundry VTT integration. We have none — if you run Foundry, our path is recording the audio separately, which takes a workaround.
• AI image generation, digital trading cards and a public campaign showcase. We do not generate images at all.
• Per-item top-ups. An extra campaign is $6 and an extra session is $2, so overflowing a plan does not force an upgrade. Our equivalent is a 5-hour bundle at $4.95 that never expires.

What we do that Archivist does not appear to

Two things, and the second matters more than it sounds.

A permanent free plan. Archivist gives every new user 30 days free with 1 campaign, 2 sessions and 5 AI images, no card required — a good trial, and it ends. Our free tier does not expire: 2.5 hours a month, forever, plus 3 bonus hours at signup. If you run a one-shot every couple of months, you may never need to pay us anything, and that is a legitimate way to use it.

A feedback loop from your own wiki. The entries your sessions create are fed back into every later transcription — the most recently updated wiki entries, the player characters, the recorded nicknames — so the spelling of an invented name settles once and then stays settled week to week. The mechanics are on the transcription page.

Neither of us labels speakers

Worth stating because it is the most common disappointment with tools in this category. Our transcripts do not say who spoke, and nothing on Archivist’s site claims per-speaker labelling either. If attribution is the feature you are shopping for, Saga20 advertises voice recognition on every plan — a different comparison, made separately.

If you need to know which player said which line, record per-speaker tracks — Craig on Discord and Zoom’s per-participant option both do it free — and keep those files to check a line against its voice. The transcript itself still comes from a normal combined recording with either tool; a recorder that writes several audio tracks into one file, like OBS, works as-is with us because the tracks are mixed automatically on upload.