Sweep, AI autocomplete for JetBrains that rewrites code
Hi HN, I've been a JetBrains user for the last couple of years, and felt that JetBrains was missing a strong autocomplete tool. Over the last 6 months, we've fine-tuned our own model, built our own inference stack, and integrated it deeply into the IDE. You can use the autocomplete plugin for free (no sign up required) on the JetBrains marketplace. Here's how we got here: We started with standard FIM (fill-in-the-middle) autocomplete. Other AI code completion tools in JetBrains can only "add" code at your cursor position. This can be helpful when writing assert statements in unit tests, but…
What it does
In the maker’s words, at launch
Hi HN, I've been a JetBrains user for the last couple of years, and felt that JetBrains was missing a strong autocomplete tool. Over the last 6 months, we've fine-tuned our own model, built our own inference stack, and integrated it deeply into the IDE. You can use the autocomplete plugin for free (no sign up required) on the JetBrains marketplace. Here's how we got here: We started with standard FIM (fill-in-the-middle) autocomplete. Other AI code completion tools in JetBrains can only "add" code at your cursor position. This can be helpful when writing assert statements in unit tests, but FIM is less useful for cases like adding a new parameter to a function. Autocomplete feels much better when it can rewrite code in addition to simply adding code. This is typically called "next-edit." To get this capability, we trained a model on granular user actions such as arrow key movements, cursor jumps, and keystroke-level data. This works really well for tasks like adding enumerate to a Python for-loop, refactoring conditionals, and other repetitive changes. Another problem is that this can be very slow. Out of the box with vLLM, each request had a median latency of 1500ms (too long). To optimize this, we rewrote TensorRT-LLM to support N-gram speculative decoding, which lets us serve completions at a median latency of 94ms. We also wrote more about it here: https://blog.sweep.dev/posts/next-edit-jetbrains Finally, to get full codebase context awareness, we actually have one unique advantage over VS Code. The JetBrains codebase index (via their Program Structure Interface) is exceptionally well-built and has already indexed the entire codebase. This means we can quickly access the definitions of arbitrary functions or classes. To get extremely precise codebase context, we fetch the definitions of code symbols around your cursor and pass them to our model. We've spent a lot of time getting the details right, and we'd love to get your thoughts and feedback!
Does the same job
all alternatives →
- CICodeMatch: IDE autocomplete on overdrive2011 · languageinterfaces.com · ▲102
- FAFast autocomplete service written in Go and JavaScript2015 · autocompeter.com · ▲76
More ai this month
the category →
I trained a 125M-parameter transformer to autocomplete piano performances in real time (~108 notes/sec on an iPhone 15). The idea is basically GitHub Copilot or Tabnine, except instead of prompting it with code, you prompt it by playing a few notes on a MIDI piano. The model then continues what you played, entirely on-device. The app is free if anyone wants to try it. Happy to answer questions about the model, training, Core ML, or the many things that didn't work.
AI · 17d ago · simedw.com
Astute▲585Automate your B2B brand going viral, with new media creators
AI · 18d ago · company-app.joinastute.com
Hey HN, Henry from Cactus here! We previously released Cactus Needle, a 14MB agentic LLM for tool call, device use, and structured extraction for phones, wearables, smart homes, small robots and microcontrollers. We got really great feedback here, and have now incorporated the suggestions to release Needle 2. The whole model is a single 14MB binary that runs a full session in 28MB of RAM; 45m parameters at 2bit compression. Needle hits 500 tokens/sec decode speed on a Raspberry Pi 5, sits between 400-1,500 tokens/sec on VR devices like Meta Quest 3S and Apple Vision Pro, and ranges…
AI · 27d ago · cactuscompute.com
Launched alongside, October 2025
the whole month →
- SA
I went down the rabbit hole on a side project and ended up building this: Strange Attractors(https://blog.shashanktomar.com/posts/strange-attractors). It’s built with three.js. Working on it reminded me of the little "maths for fun" exercises I used to do while learning programming in early days. Just trying things out, getting fascinated and geeky, and being surprised by the results. I spent way too much time on this, but it was extreme fun. My favorite part: someone pointed me to the Simone Attractor on Threads. It is a 2D attractor and I asked GPT to extrapolate it to…
AI · Oct 2025 · blog.shashanktomar.com
- ASAutism Simulator▲779
Hey all, I built this. It’s not trying to capture every autistic experience (that’d be impossible). It’s based on my own lived experience as well as that of friends on the spectrum. I'm trying to give people a feel for what masking, decision fatigue, and burnout can look like day-to-day. That’s hard to explain in words, but easier to show through choices and stats. I'm not trying to "define autism". I’ve gotten good feedback here about resilience, meds, and difficulty tuning. I’ll keep tweaking it. If even a few people walk away thinking, "ah, maybe that’s why my coworker struggles in those…
Life & fun · Oct 2025 · autism-simulator.vercel.app
