Leap – Full-stack AI developer agent that deploys to AWS
We just launched Leap, an AI developer agent that helps you build and deploy apps with real backends to your own AWS or GCP cloud. We built it because every AI app builder we tried fell short. They were fine for prototypes, but not for real systems. No support for proper backend infra, no isolated test environments, and sometimes shockingly no way to own your code. We build Leap to be more appropriate for developers: - Connect GitHub to put your code in a repo you control - Work with code generation using pull request–style revisions with diffs - Get Architecture diagrams and API docs for…
In plain words
Leap is an AI developer agent that builds and deploys full-stack applications to AWS or GCP. It connects to GitHub to keep code in user-controlled repositories and generates code through pull request-style revisions with diffs. The tool provides architecture diagrams and API documentation, includes isolated preview environments for development, and supports automated deployment. It's designed for developers building production systems rather than prototypes, offering proper backend infrastructure support and code ownership using Claude and the open-source Encore.ts framework.
written from the facts on this page · September 2026
From the sources
In the maker’s words, at launch
We just launched Leap, an AI developer agent that helps you build and deploy apps with real backends to your own AWS or GCP cloud. We built it because every AI app builder we tried fell short. They were fine for prototypes, but not for real systems. No support for proper backend infra, no isolated test environments, and sometimes shockingly no way to own your code. We build Leap to be more appropriate for developers: - Connect GitHub to put your code in a repo you control - Work with code generation using pull request–style revisions with diffs - Get Architecture diagrams and API docs for your app as you build - Develop using an isolated preview environment - Deploy either via open source tooling or the integration with Encore Cloud for automated deploys to AWS/GCP Leap uses Claude 4 Sonnet and runs on our open-source framework Encore.ts[1], which provides declarative infrastructure (like a cloud-agnostic CDK-layer). We found the framework to be a good fit for LLM code generation as both application logic and infra is defined in the same context, and the declarative nature makes it less error prone. Right now we think Leap is best for starting new projects or building new isolated services in existing larger systems. The biggest challenge right now is dealing with larger codebases as models struggle to cope with large contexts. And while we're doing a few things to optimize this, we intentionally made it so that you can switch to your IDE any time and keep working there. To run your app locally you only need the open Encore CLI installed, which also can be used to build the app as a docker container. We're just getting started and would love your feedback if you try it. And also it would be helpful to understand what would make this useful for you? - You can try it here: https://leap.new - Here's the launch blog post and video: https://leap.new/launched [1]https://github.com/encoredev/encore
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 · 16d 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 · 26d ago · cactuscompute.com
Launched alongside, May 2025
the whole month →
- C9
Life & fun · 2025 · felixrieseberg.github.io
