Subscribe
Sign in
Pieces Copilot — The AI coding sidekick that understands your workflow
Pieces Copilot helps you solve your most complex coding problems through a contextual understanding of your workflow. Leverage local, cloud and custom LLMs within our native desktop app and integrations for a secure, tailored, and productive experience.
135
Replies
Best
@tsavo_at_pieces Pieces Copilot sounds impressive for streamlining coding workflows with its AI assistant and personalized suggestions. Love the cross-platform support and focus on security with on-device data processing. Curious to know, does it support a wide range of coding languages effectively? Excited about its potential impact!
@tsavo_at_pieces how are you syncing it between the devices when it also says that "data stays on-device", not sure about that.
Otherwise, it looks like a great product, I will definitely give it a try.
Great find @kevin as always!
@tsavo_at_pieces @sichen_shao Thank you so much for the support! In terms of auto-enrichment and language classifications, we support around 50 coding languages.
@tsavo_at_pieces @kevin @arnav1712 we do not have cross-device sync yet! We're still working on that feature. As of now the only way data can be shared is through manual backup and sync, sharing to GitHub Gists, or creating shareable links for individual snippets.
Hey hey, all!
Really appreciate the the support and insights here 🙏
Definitely check out our YouTube channel and Discord Community for more details on all things considered!
youtube.com/@getpieces
discord.com/getpieces
We're so excited for this second Phase i.e. the Pieces Copilot finally going mainstream... also can't wait for the third Phase i.e. the Pieces Feed hitting everyones workflow in Q1!
Definitely give pieces a go, try out the extensions and let us know thoughts :)
Thank you for being on this journey with us 🙌
@sentry_co Glad you like this tech! There are couple differences, but the one I like the most is we let you use a model of choice, so you're not stuck with just one model, and also we let you use air-gapped on-device models so your code and questions don't leave your machine if you choose them not to :D
Also we keep working on supporting more and more models, so the selection will only grow!
@kuba_at_pieces Nice! Im no expert but I guess model memory is important when understanding more of the codebase? So the more context memory the model has the better it is to understand the codebase? What is the current highest context you support?
@sentry_co I'm not positive I fully understand your question. In terms of system memory, each model requires different amount of it. The more context you add, the more system memory you'll need. In terms of how big a context scope can be, I don't believe we set any limit on that. You can define what needs to be in the context of the conversation you're having with Pieces Copilot and the only limit will be your hardware.
@kuba_at_pieces Its more my experience the last few months. If you give the model too much contextual data it trottles it in a way.
@sentry_co We try to optimize our queries and context management to our best abilities. We can always hope for a release of an LLM project that is deeply optimized for tons of context and won't stutter under any load ;)
I am not a coder, but I have had to use minimal amounts of coding for things sometimes and I can imagine this is going to be incredibly helpful to coders out there!
@annie_chopra Thank you for checking out Pieces Copilot! I can tell you it will be helpful even to you if you want to understand some code you see for the first time! Give it a try, you'll love it 😁
@annie_chopra Hope you'll find it helpful Annie! We've built it to provide a seamless experience for all kinds and levels of developers.
You can get started with it for something simple like just making it easier for you to manage and share snippets you come across but it can do more than that and can take things up a level when you need it 🙌
Congratulations on the launch! 🎉
This is an exciting and challenging idea brought to life and I'm sure this is going to save lots of time and efforts for developers in building robust applications without sacrificing their peace. (You could consider renaming it to Peaces Copilot) as it will give a peace of mind to those devs! 😅
@moingshaikh Thank you for your comment! Not sure if we adopt the new name, but we're always open for feedback 😊
Hey @campak 👋
We don't have any generated in-line code completion, however, we do have code completion using your saved snippets. So if you have a class template saved for example, you can use code completion in VS Code to fill the rest of that class out.
That's awesome that you heard of us from a conference! Was there a talk that mentioned us or how did Pieces come up? (If you don't mind me asking)
@masnwilliams Nice, good to know! Someone who gave a talk mentioned that it was one of their favorite products to help them in their dev journey :)
@campak Oh that’s awesome! If you could remember who it was, I’d love to reach out to them but no worries if you can’t!
Enjoy the holidays!
@masnwilliams since it was a lightning talk, unfortunately I do not have a list of those names
Merry Christmas, though!
Hey @ritavdas 👋
Thanks for the support! We are focusing on developing simply because of the ecosystem we have built with Pieces for Developers.
You will see over the next year, we will expand to different sectors like Pieces for Designers and so on, but developers are the primary users right now.
That said, we do have non-developer users who enjoy our products as well so it goes both ways 👍
Pieces is awesome - our team has been using it for months. Makes managing the entire dev workflow so much easier, for both ICs and non-dev managers/collaborators. Super clean + powerful, really excited to dig into the update.
Hey @kai_mckinney 👋
Thank you for your support! We have a ton of awesome features planned for the early 2024 roadmap as well so stay tuned!
This seems like a powerful tool for coders. A suggestion would be to include a collaborative feature where developers can share and discuss code snippets or solutions within the app.
Hey @narendra_solanki 👋, thanks for the feedback and support!
As far as sharing goes, you can already share snippets with anyone by simply generating a shareable link.
Then diving more into the discussion topic, we did recently add "Annotations", which we plan to evolve into a way for users to leave comments/feedback on snippets.
The Pieces team is really excited about integrating more collaboration features so stay tuned as we roll into 2024!
@narendra_solanki Hey hi Narendra, thanx for the support 💪 Btw, we already allow you to share snippets with others. You can try it out with latest releases of any of our products. Check it out here,
https://docs.pieces.app/features...
And we're actively looking to improve Pieces products wrt to collaborative feature. Feel free to send me a dm if you've anything specific in mind on collaborative features.
Here's some more info on where we're at on collaborative features,
https://www.producthunt.com/post...
First of all, I have to say I love the design of the UI and the product video! So smooth and well done! Seems like a great tool for developers! Congrats on the launch!
@matej_cabadaj Thank you so much! We put a lot of work into details, which can be seen everywhere across our products, not only the videos ;)
Let us know what are your impressions on Pieces Copilot!
It's basically a snippet taker for code, that helps you search on it using LLM. I thought it was automatically inspecting my codebase like CoPilot.
Such a great tool for developers. This is the future of AI tools for development. A tool that can understand context of repo, help me to understand code base and can give me suggestions, all Locally 🔥
Hey @hiteshchoudhary 👋
Great to see you again and thank you for your support!
This sounds like a very useful tool!
If it really works as advertised, it seems so powerful!
Will need to take some time to test it very soon.
Congrats on the launch!
Hey @plbompard 👋
Thank you for the support! We'd love to hear back from you after you give it a good test 👍
I think you'll find that it is next level, our Copilot blows my mind every single day!
@exo_cody Continue talking to users, integrating with new tools, and shipping new features every week!
Hey there! Pieces Copilot sounds like an amazing tool for developers to manage and deeply understand the abundance of resources they interact with on a daily basis. I'm really looking forward to the launch of your product and would love to try it out once it's live. Do you have any advice for someone like me who's about to launch their own product? Also, I'd love to get your feedback once Pieces Copilot is ready. Feel free to click on the "Notify" button to get notified when it's live. Thanks in advance!
@sophiawms Our Copilot can be seen in the latest release of Pieces, feel free to check it out and let us know what you think! I've followed your product, it also looks interesting :)
@mark_eser It's free to use for now, so you can decide if it's better, we believe it is 😉
@mark_eser GitHub Copilot is more focused on code completion - Pieces Copilot is better suited for asking questions about code, generating more personalized snippets based on the context of your project, on-device security, and we have a code snippet manager built-in so you can keep track of useful workflow materials along the way!
Pieces is such an amazing product, I can't believe you get all these dev tools packed for free! Security-wise, it's just amazing, I didn't know it's possible to use LLMs like this locally, without sending anything to the cloud :D
Hey @domagoj_vidovic1 👋
It still blows my mind every day and it is only getter better day-by-day!
Thanks for your support!
@domagoj_vidovic1 Thanks so much Dom! You showcase it very well on your social vids
Hey @bayazid_malikov 👋
Thank you for the support! As far as context limits go, locally we use algorithmic snippetization and summarization to maximize 8k contexts which scales up to 128k for upcoming local and cloud models.
The users are also able to select which messages the LLM should specifically reference if things are getting off-course.
@alex_gavril1 Appreciate the support Alex🙏 Feel free to try it out and let us know how you find Pieces Copilot and any places where we can improve🙌
Pieces for Developers