Privacy Policy

Who we are

Selendis ("we") makes a collaborative, AI-native video editor. This policy covers selendis.com, the Selendis waitlist, and the Selendis desktop application. Contact: gavin@selendis.com.

What we collect

When you join the waitlist:

• Google sign-in — your email address, name, and verified-email status. We use Google only to identify you; we do not receive your password.
• Linked social accounts (optional) — if you connect X, YouTube, or Discord we store your public account identity (ID, handle, display name), public audience statistics (e.g. follower or subscriber counts), and the OAuth tokens needed to verify the connection. Tokens are encrypted at rest.
• Referrals — which invite link brought you here, and which signups your own link brought in.

How we use it

• To hold your place, compute your waitlist rank, and award points.
• To verify follow/subscribe/join tasks you choose to complete.
• To review waitlist members and send access invitations.
• To contact you about your waitlist status and product access.

We do not sell your data. Your email and social statistics are never displayed publicly; the public leaderboard shows ranks only, and your username appears on a share card only when you share your own link.

The desktop application

When you use the Selendis editor:

• Account & projects — your account identity, project metadata, and (while cloud sync is on) the media you import, stored in our cloud so projects follow you across devices and collaborators can join. With cloud sync off, your media stays on your device.
• Editing preferences — the settings you give the editor and its assistant (for example, your default effect values and the words you tell it to cut) are stored with your account so they follow you across devices.
• Usage data — which features you use and how the app behaves (navigation, errors, performance), so we can improve the product.
• Editing activity — the editor records your editing sessions (timeline changes, tool use, interactions with the AI assistant) tied to your account. We use this to operate features like undo and session review and to improve Selendis.
• Deleted projects — when you delete a project whose edits were used to build training examples (with your consent), the session recording and a reduced-resolution copy of the media, plus a small number of full-resolution still frames from the parts used in the edit, are kept with your account for training and quality review until you delete your account; deleting your account removes them.
• AI features — messages you send the AI assistant, and the project content needed to act on them, are processed by our AI providers to deliver the feature. Voice transcription runs on your device by default; when local transcription isn't available, the audio is processed on our servers instead.
• Your other projects, when drafting a script — when you ask the AI assistant to draft a script, it can use your earlier work as a style reference so the draft sounds like you: scripts you wrote, pasted, or edited yourself in other projects you created, and the transcribed speech from finished videos in other projects you currently own or edit. This happens only while everyone who currently owns or edits the source project has opted in to model training below, and only after we remove the names of that project's members who have opted in. Your scripts are used only for your own drafts. A shared project's transcribed speech can also inform the drafts of anyone else who currently owns or edits it; to stop that, turn off model training. The material is sent to our AI providers with your request, like other assistant requests; using it as a reference does not by itself add anything to a training dataset, and the draft it helps write is treated like any other draft. To stop your own drafts using any of this, turn off “Use my editing history to personalize the agent” in Settings.
• Your Library videos, when drafting a script — if you have contributed finished videos to your Library and asked for them to be used for your script voice, your own lines from their transcripts (never anyone else's) may be used as examples when the AI assistant drafts scripts for you, so the draft follows your style, length and series. They are sent to the AI provider that writes the draft, with your request. This is personalization, not training: it is controlled by “Use my editing history to personalize the agent” in Settings (on by default), and using them as examples does not by itself add anything to a training dataset. If you also let Selendis look at your Library videos, an AI reads each video's title (and the start of its transcript) to name what kind of video it is, and looks at one frame for each sentence spoken in it to note how it is shot. Those frames, titles and your own lines are processed by AI providers running openly published AI models to write scripts in your style, and — only while model training below is on — to improve Selendis's AI.
• Connected accounts — as on the waitlist, plus performance statistics for your published videos (e.g. views and audience retention over time), used to personalize editing suggestions, to measure how edits perform, and — only if you opt in to training below — as training data.
• Collaboration — your name, username, and live presence (cursor, selection) are visible to other people in the same project. In a shared project, your editing activity there (in structured form — for example, which clip you replaced or how you adjusted a placement) also informs the AI assistant for everyone working in that project, so the assistant can learn the project's direction from all of its editors. This applies only within projects you have joined, and is separate from model training below — it does not add your data to any training dataset.

Model training (opt-in)

Selendis learns from real editing decisions — but only from people who choose to contribute. Training contribution is off by default. If you turn it on in Settings, your editing decisions (for example, how you cut, arrange, and refine a project, and before/after examples from your sessions) and how the videos you publish from Selendis perform with viewers (for example, views and audience retention over time) may be included in training datasets under a pseudonymous ID, and you earn contribution rewards. When training is on, your sessions, including your edits and the assistant's suggestions, may be reviewed and used to train and improve Selendis's AI. If you have also contributed finished videos to your Library and asked for them to be used for your script voice, the transcripts of those videos — your own speech only, never anyone else's — may train your own script writing and, while this setting is on and under the Library agreement you accepted, Selendis's shared models that help every creator. While this setting is on, the notes and labels Selendis makes when it looks at your Library videos, and any label you change, may also be used to improve Selendis's AI; notes and labels made from videos you add to a project as references are never used for training. Your email and account identity are not part of the training data. In shared projects, each collaborator's own setting governs their own editing contributions, with one exception: comments. A project's comments belong to the project, so if the project's owner has turned contribution on, the comments in that project and how they were resolved (including comments written by collaborators who have not made a choice about contribution) may be included in training datasets under pseudonymous IDs, with people's names removed. If you have turned contribution off, declined it when asked, or chosen "Don't use my comments in shared projects" in Settings, your comments are never included, in any project. You can turn contribution off at any time — future edits stop being contributed immediately. If you turn contribution off or delete your account, we stop using your contributions for new training, and any model trained on a dataset that included them is retrained without them or retired within 30 days.

Security logs

Across the site and app, sign-ins and administrative actions (such as account or username enforcement under our terms) are logged with IP address and device information, for abuse prevention and account security.

Google API Services

Selendis's use and transfer of information received from Google APIs adheres to the Google API Services User Data Policy, including the Limited Use requirements. YouTube data — your channel identity, subscriber count, and whether you subscribe to the Selendis channel, and, if you connect your channel in the app, analytics for the videos you publish (such as views and audience retention) — is used only for the purposes described above: the waitlist, personalizing editing suggestions and measuring how edits perform, and, only if you opt in to model training, as training data for Selendis's editing models.

Retention & deletion

Waitlist data is kept while the waitlist operates and, if you're admitted, becomes part of your Selendis account. Deleting a project in the app removes it from active storage. You can disconnect a linked social account at any time, and you can request deletion of your waitlist or account data by emailing gavin@selendis.com from your sign-up address — we'll action it within 30 days.

Changes

We'll update this page if our practices change; material changes will be noted at the top with a new date.