Skip to the page
Look
Try the beta

14 November 2026 · Version 2.12.0

Terms of use

These are the terms on which you may use Selver. They are meant to be read. If you do not agree with them, do not use Selver.

Contents

  1. What Selver is
  2. Who can use it
  3. Before you record
  4. What Selver does with a recording
  5. What a reflection is, and is not
  6. What is yours and what is Selver’s
  7. Using Selver fairly
  8. While Selver is unfinished
  9. If something goes wrong
  10. Ending
  11. Changes to these terms
  12. The law that applies
  13. Contact
  14. Credits

1What Selver is

Selver is an app that listens to a conversation when you ask it to and afterwards shows you your own side of it: how much of the floor you held, where you cut in, what you said that worked, and, when there is one, one thing to try next time. When it listens to you alone (“Only me”), it shows only what your own side can tell: not how much of the floor you held, or where you cut in.

Selver is unfinished and it will change. Everything it does on your phone is free, for good: Before, a word to yourself, the lines you keep, the page that shows what Selver holds, and the reflections you already have and their walkthroughs. A plan is for new reflections, and your first three reflections are free. Until you are told otherwise, nobody is charged for a plan.

2Who can use it

You must be 18 or older.

You need an account, and you are responsible for what is done with it. Keep your password to yourself.

3Before you record

Selver records other people as well as you. Before every recording, you must tell everyone who will be heard that the conversation is being recorded and that Selver will transcribe and analyse it. Selver asks you to confirm this before it starts, and keeps a record that you did.

When you choose “Only me”, Selver asks you instead to confirm that nobody else can be heard, and keeps a record that you said so until you delete that reflection, every reflection or your account. Selver relies on your word for this: it cannot always tell another voice from yours. Say it only when it is true. If anyone else can be heard, choose “Both of us” and tell them, as for any recording.

In many places the law asks for more than telling people: it asks for each person’s consent. You are responsible for knowing and following the law where you are, and where the people you are talking to are. If you are not sure, ask them, and record only if they agree.

Do not record anyone who has not been told, anyone who has said no, or a conversation you have no right to record.

4What Selver does with a recording

The audio stays on your phone, where the voices are told apart and what was said is written down. That text is sent to Selver’s service and analysed there. Parts of that work are done by other companies, which the Privacy policy names. By recording, you ask Selver to do this, and you confirm that you have the right to ask.

What is kept, for how long, and what you can delete are set out in the Privacy policy, which is part of these terms.

5What a reflection is, and is not

What Selver counts from a recording, such as how long each person spoke, comes from automatic analysis of the audio. It can be wrong: it can mistake who is speaking, or mishear a word.

The sentences, the suggested move and the estimates in a reflection are written by a language model. They are estimates, not a validated assessment, and Selver labels them as estimates.

Selver is not a medical device or a health service. It does not diagnose, treat or assess any condition, and nothing it shows is medical, psychological or legal advice.

Selver is about your side of a conversation. Do not use it to assess, monitor or make decisions about another person.

6What is yours and what is Selver’s

The recordings you make and the reflections made from them are yours. You give Selver permission to process what your phone sends it and to store your reflections, including through the companies named in the Privacy policy, only in order to provide Selver to you.

The app, its design and its name belong to Selver. These terms let you use the app for yourself. They do not give you ownership of any part of it, and you may not copy, resell or take it apart.

7Using Selver fairly

Do not use Selver to break the law, do not try to reach other people’s data, and do not try to disrupt the service.

8While Selver is unfinished

Selver may change, pause or stop, and a reflection you have may stop being available. Selver is provided as it is, with no promise that it is accurate, uninterrupted or fit for a particular purpose.

If new reflections come to need a plan you pay for, you will be told in the app before you are asked to pay for anything. Everything else Selver does on your phone stays free, and nothing you have already made is ever held back for want of a plan.

9If something goes wrong

As far as the law allows, Selver is not liable for indirect or consequential loss arising from your use of it, or from your not being able to use it.

Nothing in these terms limits liability that the law does not allow to be limited, or takes away rights you have as a consumer where you live.

10Ending

You can stop using Selver at any time, and you can delete your account in the app, under You, in “Account”. The Privacy policy says what that removes.

Selver may suspend or close an account that breaks these terms, in particular one used to record people who have not been told.

11Changes to these terms

When these terms change, the date and the version at the top change with them. If a change affects how you use Selver, Selver will tell you in the app before it takes effect.

12The law that applies

These terms do not yet name a governing law or a court. Selver will name them before it is released to everyone. Until then, the law that applies is the law that would apply with no choice made, and nothing here limits rights you have under the law where you live.

13Contact

Questions about these terms go to kendara2023@gmail.com.

14Credits

Your phone listens with speech models that others made and published under the Creative Commons Attribution 4.0 licence (https://creativecommons.org/licenses/by/4.0/). The model that writes speech down is Parakeet TDT 0.6b v3, by NVIDIA. The models that tell voices apart are the community-1 pipeline, by pyannote, with speaker embeddings from WeSpeaker and parameters from BUT Speech@FIT.

Selver speaks with Kokoro-82M, by hexgrad, in its voice af_heart, in a conversion for Apple’s Core ML by laishere and Fluid Inference, with the pronunciation dictionary of misaki, by hexgrad, and a pronunciation model by PeterReid, all published under the Apache License 2.0.

Selver uses them as Fluid Inference converted them for Apple’s Core ML, so the files on your phone are modified versions of the originals, not the originals themselves. Fluid Inference publishes the conversions at https://huggingface.co/FluidInference.

The app hears and speaks with FluidAudio, by Fluid Inference, used under the Apache License 2.0.

Look
Privacy Terms For the other person kendara2023@gmail.com

© 2026 Selver