Pieces Long-Term Memory Agent - The first AI that remembers everything you work on

@jackross Exactly! LTM-2 is the major upgrade in AI assistants that we have all been waiting for. Thank you for the support Jack!

@sam_parks_at_pieces Workstream Activities are a game-changer for sure 😎

@zongze_x thanks for the great technical question, you are spot on, the issue of memory contamination is complex and a core challenge in designing features like this one. I suspect you can appreciate that writing a full answer here is tough - but would make an excellent topic for a technical article (watch this space). At a high-level, our approach to identifying and minimizing contamination happens at three levels:

• On entry: we are very selective about what is added to the LTM. By analysing where the users focus is and how what they are focusing on currently relates to the big picture of there workstream we can prevent a lot of corruption at source.

• On roll up: when we roll up memories into periodic summaries our agent looks for narratives and themes across workflow elements. When we find contradictions, we resolve them by comparing those narratives to cut out random chatter and keep focus on core tasks.

• At query time: when you interact with your workstream data, through the copilot or the summaries, those interactions are used to infer which aspects are useful/truthful and which are not, which allows us to elevate quality information whilst demoting the noise.
  

Additionally, signals from all of these levels are used to periodically clean contamination from your stored memories. It's a work in progress but I have found the LTM to be much more resistant to context corruption than other solutions out there.

@themisty Hey Krishna! Thank you for checking out the launch! Your paper sounds really interesting, I'm sure a lot of people from my team would enjoy reading it. Where will it be published?

@elliezub Dear Ellie, I am excited to have readers interested already :) it will be on Arxiv, fingers crossed. Still under heavy reviewing lol.

@themisty Sounds great! Looks like we are connected on Linkedin now, so hopefully you will post about it once it's published. Can't wait to read it!

@themisty Can't wait to read your paper Krishna! Thanks for the support as always!!

@elliezub Thanks will definitely ping the team. In the meantime I definitely don't mind if you guys can feel my product nonilion ? I don't mind a solid feedback 🙌

Hey @tleyden, Thanks for your support! Means a lot;

PiecesOS captures your information and works with couchbase database locally to store this information. Plugin's allow you to bring this memory in your choice of IDE. I am usually coding in VS Code and chrome so I have both the plugins.

While using the plugin you can provide more context by adding the codebase etc. Hope this clarifies your question!;

@tsavo_at_pieces @tleyden you don't need to install the plugins for memory capture, pieces runs just like magic, still the VS extension makes working with pieces easier, so I would install it anyways.

Hey @tleyden , thanks for the great question and your awesome intuition—you’re spot on!

With LTM-2 enabled, our system already leverages a uniform screen segmentation (we never actually record any video.. too heavy to process 😅), vision processing, and accessibility APIs, so it works incredibly well out of the box. That said, installing plugins (like the VS Code one) can provide even richer data—think deep stack traces, AST details, and more discrete file paths.

Believe it or not, we're already leveraging some integrations sending what we call “Tier 3” data, which gets blended with the lower-level visual and accessibility data and interconnected there on the device through a couple classic algos. That said, both data sources direct-from-plugin and uniform at the OS are unique and additive so you'll definitely be seeing us continue to invest in the plugins 🌟

Anyway, hopefully that answers your question and thanks again for your support!

Cheers,
-Tsavo

@bishoy_hany1 Same! The possibilities are really endless with how it can improve your workflow. Thank you for the support Bishoy!

@henry_habib Thank you very much, do try it out and let us know your favorite prompt;

@henry_habib Thanks for all the support Henry! Can't wait to see how LTM changes the way you use AI!

 @whatshivamdo Really appreciate the support Shivam! LTM is definitely the biggest productivity boost I've gotten from LLMs in a loooong time!

 @whatshivamdo Thank you for the support Shivam! 🤝

@terrence_kelleman That's a great question. Currently we are focusing on tech professionals such as developers, support, devops and so on, but the technology can and most likely will be adapted for all knowledge workers. Will it work for elder care or consumers in general? That's a great question. Because it is intrinsically tied to the activities you are doing on your computer it can help with this if it is related to them using a computer. If these folks are using a desktop or laptop computer to plan activities, get medical updates, and so on, then Pieces can summarize these. "When am I meeting Susan for lunch this week?" "What was the link my doctor shared on best practices with this medication?".

Thank you for your support, @terrence_kelleman. I have noticed that people from various domains are using Longterm Memory for their specific use cases. Personally, I utilize it to manage my schedule since I deliver many workshops and communicate through multiple channels. Pieces helps me easily find available time slots.

I can see how this could be beneficial for many others in numerous ways! Do share your feedback if you are using it for a different usecase;

@stemonte I've not tried that large, I "only" have a 39" ultra wide, but it works fine for me! If you want to lend me your setup for a while I'll be happy to test for you 😁

@stemonte Be aware that the larger the monitor, the more system resources will be used. But Pieces uses only 1-2% of CPU typically, so the increase will be minimal. We only extract memories from the current active window as well to keep them more human-centric, so using multiple monitors has no impact. I'm also guessing that if you have large monitors you probably are not running on a 2016 low spec IBM Thinkpad 😝. So your CPU impact will be negligible.