Narrative recaps
Every completed session becomes a readable, novel-style chronicle with the drama, discoveries and finest party blunders intact.
The digital bard for your table
Lestapenna turns your tabletop sessions on Discord into narrative recaps and a living archive of characters, places, quests and unforgettable choices.
Free, forever. No subscription, no credits, nothing to buy — buy me a coffee if it helps your table.
Already play at a table using the bot? Log in above. Not onboarded yet? Request early access.
Not only Discord
The bot lives in the voice channel, but the campaign is read on the web. Every session becomes a diary you turn page by page, next to cards for NPCs, places, quests, inventory and bestiary — and a chat with the Bard about anything the table has lived through.
The story, after the session
Lestapenna gives the table a shared memory without turning the session into admin work. The adventure stays in Discord; its history becomes easy to revisit.
Every completed session becomes a readable, novel-style chronicle with the drama, discoveries and finest party blunders intact.
Characters, NPCs, places, factions, quests, loot and timeline evolve after each session instead of disappearing into scattered notes.
Ask what happened, who made a promise or where an artefact appeared. Answers are grounded in the history of your own campaign.
Browse sessions and world data, choose your Discord server and campaign, and keep your keys, models and estimated costs visible in one place.
How it works
A deliberate four-step flow keeps the table in control—from starting the recording to reviewing the finished chronicle.
During the private beta, we help configure the Discord server and campaign.
The DM starts recording only after the group is ready and informed.
Lestapenna captures each participant and processes the session when it ends.
Read the recap and explore the updated campaign archive from Discord or the web.
Recording with intent
Recording is a deliberate session action, never a background feature. Server administrators remain responsible for informing participants and obtaining the consent required for their table and jurisdiction.
Free, and open source
Lestapenna is a personal effort donated to the tabletop community. There is no subscription, no credits, no payment of any kind: the software is free and its source is public under AGPL-3.0. The AI runs on your own provider keys, so what it costs is charged straight to your account — and you can see the estimate before every action.
Bring your own key
OpenAI or Google Gemini
Or your own machine
Ollama and your own PC
Self-host it
Docker, in one command
Shared public bot: the free hardware currently supports at most two simultaneous recordings across two Discord servers. For dedicated capacity and smoother continuity, self-host your own instance; donations help us invest in shared cloud infrastructure.
What we do not do: we receive no money, we sell no data, and there is no analytics or tracking script anywhere in this project. Third-party services are named explicitly in the privacy policy: Discord, Oracle Cloud (Milan) for storage, and the AI provider you choose — where the relationship is directly between you and them.
Support the project
Lestapenna asks for a donation instead of a subscription, and the difference is the whole point: every feature is free for everyone, always, and giving unlocks nothing. If it is useful to your table, this is what keeps the shared bot running.
Entirely optional. Nothing is locked behind it, and there is nothing to cancel.
Questions from the tavern
We are opening access gradually while the authenticated web dashboard and Discord publication flow are completed.
Nothing to us. The AI runs on your provider key, so you pay OpenAI or Google directly — a four-hour session is typically a couple of euros, and the estimate is shown before you start. If you run transcription and models on your own machine, it costs nothing at all.
Nobody. There is no company, no revenue and no investor behind Lestapenna: it is one person's project, released under AGPL-3.0 so it stays open for whoever comes next. Sponsorship on GitHub is welcome but buys no feature.
Yes. The source is public and the whole stack runs with Docker. On your own server nothing leaves your machine except the calls to the AI provider you configure.
What a session needs: the audio you explicitly record, its transcript, and the campaign material generated from it. No analytics, no tracking, no profiling — and nothing is ever sold or shared. The privacy policy lists every third party involved.
Private beta
Tell us about your table and preferred language. We will get in touch as private beta capacity becomes available.
Request early access