Terms of Service

Last updated: 2 September 2026

1. The service

Lestapenna ("the Bot") records tabletop role-playing sessions held in Discord voice channels, transcribes them, and generates AI-powered summaries and campaign archives. By adding the Bot to a Discord server or using its commands, you accept these terms.

The service is free and we receive no payment of any kind for it. There is no subscription, no credit, no purchase and no billing relationship between us: nothing you do here results in an invoice from us, now or later.

1a. Open source

Lestapenna is released under the GNU Affero General Public License v3.0, and its source is public. You may use, study, modify and redistribute it. The AGPL adds one condition that matters here: if you run a modified version as a network service for others, you must offer those users the corresponding source of your version. Nothing in these terms restricts the rights that licence grants you; where the two disagree, the licence prevails.

2. Eligibility and responsibility

• You must be at least 13 years old, or older if the law of your country sets a higher minimum age, both to use the Bot and to take part in a recorded session. This mirrors Discord's own Terms of Service, and it applies to everyone in the voice channel, not only to whoever runs the commands.
• You must comply with Discord's own Terms of Service and Community Guidelines.
• Server administrators are responsible for obtaining participants' consent to recording, as described in the Privacy Policy.
• You may not use the Bot to record people without their knowledge or for any unlawful purpose.

2a. Reporting a problem

If something is broken, or if the Bot is being misused on a server, tell us: $report ($segnala) from Discord, the red button in the web app, or the address in section 9. Reports are read and acted on.

3. AI keys and what they cost you

• You bring your own AI keys. Transcription, summaries and the campaign archive run on accounts you hold with providers of your choice — currently OpenAI, Google, Anthropic or Ollama Cloud — or on hardware you own. We never resell provider capacity and never take a margin on it.
• Those costs are yours and are billed to you directly by that provider, under your own contract with them. We are not a party to it. What we do is show an estimate before an action and the real consumption after it; estimates are approximations and are not a price, a quote, or a guarantee.
• Keys you entrust to the instance are encrypted at rest and are never returned by any endpoint. You can replace or delete them at any time. Their validity, their credit and their spending limits remain your responsibility.
• You bring your own storage too. The recordings and the pictures your table generates go to a bucket on an account you hold, with the provider and in the region you choose. The instance keeps no copy of them. A free plan is enough for a table, and nothing can be recorded until you have connected one — there is no shared storage to fall back on. The credentials, like the AI keys, are encrypted at rest and never returned by any endpoint; their validity, their capacity and their bill remain yours.
• If you self-host, no key of yours reaches us at all.
• Voluntary sponsorship through GitHub Sponsors is a donation to the author's work. It buys no feature, no priority and no service level: everything is free for everyone regardless.

3a. Who is responsible for what (GDPR)

When a session is recorded, the server administrator who runs it is the data controller: they decide to record, why, and what to do with the result. Lestapenna acts as a processor on their instructions, and the AI provider is a sub-processor under their key. This is not a way of passing the buck — it is the correct allocation, and it comes with obligations on our side:

• Purpose limitation. Audio and transcripts are processed only to produce the transcript, the recap and the campaign archive of that table. They are never used to train models, never sold, never shared for anyone's marketing.
• Retention is yours to set, because the storage is yours. By default raw audio is deleted automatically after 30 days, or sooner once the session archive exists. A server administrator may lengthen that period, or switch the automatic cleanup off altogether — in which case nothing is removed for you, the recordings of your players' voices stay on your bucket until you delete them, and keeping them no longer than you need them becomes your duty as controller rather than a default we apply on your behalf. The choice is recorded with who made it and when. Transcripts and campaign data are kept while the server keeps using the Bot, so its archive keeps working.
• Sub-processors are listed in full in the Privacy Policy. Where you self-host, most of them disappear.
• Security. Provider keys are encrypted at rest with per-tenant derived keys; no endpoint ever returns their value.
• Deletion. Anyone at the table can erase their own data with $forgetme, and get a copy of it with $mydata, without asking us or anyone else. An administrator can erase the whole server with $eraseserver. Removing the Bot from a server erases that server's data automatically — recordings, transcripts, recaps and every stored file.

What remains yours to do: tell the people at your table that the session is being recorded, let anyone who does not want to take part step out, and — since the bucket is yours — keep what is on it no longer than you need it. The Bot makes this as visible as it can — it marks its own name while recording, refuses to record at all if it cannot display that marker, and posts a notice in the channel naming the commands to get a copy of your data or erase it — but it cannot obtain consent on your behalf.

4. AI-generated content

Summaries, transcripts, and other generated content may contain errors. The content is derived from your own sessions and belongs to you; we claim no ownership over your campaign material. You are responsible for reviewing generated content before relying on it.

4a. Pictures you upload

You can attach pictures to a campaign: a portrait on a character's sheet, a reference image that tells the AI what a place or a piece of armour looks like. Three things follow from that, and the first is the one that matters.

• Upload only what you have the right to upload. Your own work, something you hold a licence for, or something whose licence permits this use. Art found through a search engine is almost never any of those. By uploading you state that you have that right, and that the picture does not infringe anyone else's.
• It stays yours. We claim no ownership over it, exactly as with the rest of your campaign material. What you give us is a narrow permission — to store it, convert it to WebP, show it back to the members of your own table, and, only for the pictures you explicitly tick as references, send it to the image model your table has configured. Nothing else: it is never published, never shared with anyone outside your table, never used to train anything, and never used to promote this project. The permission ends when the picture does — delete it and it is gone from our storage too.
• What must not go up: anything unlawful; material that infringes somebody's copyright, trademark or image rights; photographs of people who have not agreed to be there.

Nobody looks at your pictures before they are stored. There is no prior screening and no automated scanning — uploads are not inspected, and we have no general obligation to monitor them. That is the honest description of how this works, and it is the reason section 4b exists: we cannot know that something should not be there until somebody tells us, and when somebody does, we act.

You can delete any picture yourself from the sheet it is attached to. Everything you uploaded also goes with $forgetme, with the campaign, and with the server when the Bot is removed — from the database and from the file storage alike.

If a third party brings a claim against us because of material you uploaded, you are responsible for it and will hold us harmless, to the extent the applicable law allows — and where your local consumer law says otherwise, that law wins.

4b. Reporting infringing or illegal content

If something stored here infringes your rights or is otherwise unlawful, tell us and we will look at it. Write to info@lestapenna.quest, and see the notice page for the full procedure.

So that a report can be acted on rather than merely received, it should contain:

• a reasoned explanation of why the material is unlawful — for a copyright claim, what the work is and what your right in it is;
• the exact location: the page address, or the identifier of the picture as it appears in the application;
• your name and email address, so we can reply and so the decision can be justified;
• a statement that, to the best of your knowledge, the report is accurate and complete and made in good faith.

We confirm receipt, decide without undue delay, and remove or disable access to material we find unlawful. Whoever uploaded it is told what was removed and why, and can reply — a mistaken report should be correctable, and a removal nobody can contest is its own kind of problem. Someone who repeatedly uploads infringing material loses access under section 6.

Until such a notice reaches us we have no knowledge of what a given picture is, and we are not liable for it — the position the Digital Services Act (Regulation (EU) 2022/2065, Article 6) gives a hosting service. Acting promptly once we do know is the other half of that arrangement, and it is the half we owe you.

5. Availability and limitations

The service is provided "as is" and "as available", without warranty of any kind, express or implied, including any warranty of merchantability, fitness for a particular purpose or non-infringement. This is a personal project offered free of charge: there is no service level, no uptime commitment and no support obligation. Processing times vary with load, features may change or disappear, and the public instance may be discontinued entirely. If continuity matters to your table, self-host it — the source and the instructions are public precisely for that reason.

6. Termination

You can stop using the service at any time by removing the Bot from your server. We may suspend or terminate access for violations of these terms or abusive usage patterns.

7. Liability

To the maximum extent permitted by law, we are not liable for any indirect, incidental or consequential damage, nor for lost data, lost recordings, or costs you incur with an AI provider — including costs caused by a misconfiguration, a key left active, or an action you did not intend to start. Since the service is provided free of charge, no monetary cap can be expressed as a fraction of what you paid us: you have paid us nothing. Nothing here limits liability that cannot be limited by law, including for wilful misconduct or gross negligence.

8. Changes

We may update these terms; material changes will be announced via the Bot or the website. Continued use after changes constitutes acceptance.

9. Contact

Operator of Lestapenna — info@lestapenna.quest