# Record your screens & mics 24/7 and connect it to LLMs

**URL:** <https://community.openai.com/t/record-your-screens-mics-24-7-and-connect-it-to-llms/876195>\
**Category:** Community\
**Tags:** project\
**Created:** [July 22, 2024, 3:31pm UTC](https://community.openai.com/t/record-your-screens-mics-24-7-and-connect-it-to-llms/876195 "2024-07-22T15:31:36Z")\
**Posts on this page:** 5\
**Page:** 1

<div class="post-metadata">

**Author:** ![louis030195](https://sea2.discourse-cdn.com/openai1/user_avatar/community.openai.com/louis030195/32/36792_2.png) [@louis030195](https://community.openai.com/u/louis030195)\
**Post date:** [July 22, 2024, 3:31pm UTC](https://community.openai.com/t/record-your-screens-mics-24-7-and-connect-it-to-llms/876195/1 "2024-07-22T15:31:36Z")

</div>

[![](https://us1.discourse-cdn.com/openai1/original/4X/f/e/3/fe364eef34a8ccf51401a78815aabd54632c0e09.jpeg "A personalized AI powered by everything you’ve seen, said, or heard. #ai #chatgpt") ](https://www.youtube.com/watch?v=0mxMGpo1anc)

Hi guys, building an OSS library to build personalized AI powered by what you’ve seen, said, or heard.

⭐ [GitHub - louis030195/screen-pipe: Library to build personalized AI powered by what you've seen, said, or heard. Works with Ollama. Alternative to Rewind.ai. Open. Secure. You own your data. Rust.](https://github.com/louis030195/screen-pipe) ⭐

It works with **Ollama** , **OpenAI** , Google, etc. Alternative to **Rewind.ai**.  
**Open source** , secure, **local** , **encrypted** on your computer.  
You own your data, written in Rust.

Also we’ve got examples to use it with **vercel/ai** , **Obsidian** , **Apple Shortcut** , etc.

Looking for feedback!

Happy OpenAI building 🚀

---

<div class="post-metadata">

**Author:** ![grandell1234](https://sea2.discourse-cdn.com/openai1/user_avatar/community.openai.com/grandell1234/32/286916_2.png) [@grandell1234](https://community.openai.com/u/grandell1234)\
**Post date:** [July 22, 2024, 5:10pm UTC](https://community.openai.com/t/record-your-screens-mics-24-7-and-connect-it-to-llms/876195/2 "2024-07-22T17:10:04Z")

</div>

Are there any benefits to using this versus Rewind?

I like that it’s open-source so I know my information isn’t going to be stolen as Rewind does record all the time.

Cool project, and thanks for making it with the project tag, this makes it easy to find in the future.

---

<div class="post-metadata">

**Author:** ![matthew.ddy](https://sea2.discourse-cdn.com/openai1/user_avatar/community.openai.com/matthew.ddy/32/427911_2.png) [@matthew.ddy](https://community.openai.com/u/matthew.ddy)\
**Post date:** [July 22, 2024, 8:39pm UTC](https://community.openai.com/t/record-your-screens-mics-24-7-and-connect-it-to-llms/876195/3 "2024-07-22T20:39:38Z")

</div>

I think Rewind is more like end-consumer app, while ScreenPipe is made for developers who want to build on top of users’ screen capture and audio data.

Instead of spending two weeks building up your own data pipeline, it’s better to get started with a reliable open source library. It’s not yet 100% reliable, but it works end-to-end!

---

<div class="post-metadata">

**Author:** ![louis030195](https://sea2.discourse-cdn.com/openai1/user_avatar/community.openai.com/louis030195/32/36792_2.png) [@louis030195](https://community.openai.com/u/louis030195)\
**Post date:** [July 23, 2024, 6:38am UTC](https://community.openai.com/t/record-your-screens-mics-24-7-and-connect-it-to-llms/876195/4 "2024-07-23T06:38:18Z")

</div>

Exactly. Basically the hard part of rewind is not the UI, the UI can be built in 1-2 hours.

The hard part is data & infra. Screenpipe abstracts this so you can easily, low code/no code, connect LLM “pipes” to your data to do things like:

- Productivity: AI PKM
- Sales: AI CRM (fill your CRM automatically while you do sales)
- Productivity: AI meeting summaries
- Productivity: Time tracking
- Engineering: Write documentations automatically
- Productivity: RAG through all your memory
- Engineering: Share how you solve problems with junior engineers
- Engineering: Share reports of what you’re working on automatically
- Freelance: More precisely track work
- Product: Aggregate multiplayer customer conversations into company’s PKM

etc.  
Open source, dev friendly, local

---

<div class="post-metadata">

**Author:** ![louis030195](https://sea2.discourse-cdn.com/openai1/user_avatar/community.openai.com/louis030195/32/36792_2.png) [@louis030195](https://community.openai.com/u/louis030195)\
**Post date:** [August 3, 2024, 1:35pm UTC](https://community.openai.com/t/record-your-screens-mics-24-7-and-connect-it-to-llms/876195/5 "2024-08-03T13:35:11Z")

</div>

What did you try yet with screenpipe ?
