←back to thread

102 points wjsekfghks | 1 comments | | HN request time: 0.001s | source

Hi, I’m Mikkel and I’m building Async, an open-sourced developer tool that combines AI coding with task management and code review.

What Async does:

  - Automatically researches coding tasks, asks clarifying questions, then executes code changes in the cloud
  - Breaks work into reviewable subtasks with stack diffs for easier code review
  - Handles the full workflow from issue to merged PR without leaving the app
Demo here: https://youtu.be/98k42b8GF4s?si=Azf3FIWAbpsXxk3_

I’ve been working as a developer for over a decade now. I’ve tried all sorts of AI tools out there including Cline, Cursor, Claude Code, Kiro and more. All are pretty amazing for bootstrapping new projects. But most of my work is iterating on existing codebases where I can't break things, and that's where the magic breaks down. None of these tools work well on mature codebases.

The problems I kept running into:

  - I'm lazy. My Claude Code workflow became: throw a vague prompt like "turn issues into tasks in Github webhook," let it run something wrong, then iterate until I realize I could've just coded it myself. Claude Code's docs say to plan first, but it's not enforced and I can't force myself to do it.
  - Context switching hell. I started using Claude Code asynchronously - give it edit permissions, let it run, alt-tab to work on something else, then come back later to review. But when I return, I need to reconcile what the task was about, context switch back, and iterate. The mental overhead kills any productivity gains.
  - Tracking sucks. I use Apple Notes with bullet points to track tasks, but it's messy. Just like many other developers, I hate PM tools but need some way to stay organized without the bloat.
  - Review bottleneck. I've never shipped Claude Code output without fixes, at minimum stylistic changes (why does it always add comments even when I tell it not to?). The review/test cycle caps me at maybe 3 concurrent tasks.
So I built Async:

  - Forces upfront planning, always asks clarifying questions and requires confirmation before executing
  - Simple task tracking that imports Github issues automatically (other integration coming soon!)
  - Executes in the cloud, breaks work into subtasks, creates commits, opens PRs
  - Built-in code review with stacked diffs - comment and iterate without leaving the app
  - Works on desktop and mobile
It works by using a lightweight research agent to scope out tasks and come up with requirements and clarifying questions as needed (e.g., "fix the truncation issue" - "Would you like a tooltip on hover?"). After you confirm requirements, it executes the task by breaking it down into subtasks and then working commit by commit. It uses a mix of Gemini and Claude Code internally and runs all changes in the background in the cloud.

You've probably seen tools that do pieces of this, but I think it makes sense as one integrated workflow.

This isn't for vibe coders. I'm building a tool that I can use in my day-to-day work. Async is for experienced developers who know their codebases and products deeply. The goal is to make Async the last tool developers need to build something great. Still early and I'm iterating quickly. Would love to know what you think.

P.S. My cofounder loves light mode, I only use dark mode. I won the argument so our tool only supports dark mode. Thumbs up if you agree with me.

Show context
brainless ◴[] No.45034954[source]
I love your video, it is very clear. I am building in this space so I am very curious and happy about all the products coming in to help the current tooling gap. What is not clear to me is how Async works, is it all local or a mix of local or cloud since I see "executes in cloud" but then I see a download-able app.

I see a lot of information on API endpoints in the README. Perhaps that is not so critical to getting started. Perhaps a `Getting Started` would help, explaining what is the desktop app and what goes into cloud.

I have been hosting online sessions for Claude Code. I have 100+ guests for my session this Friday. And after "vibe coding" full time for a few months, I am building https://github.com/brainless/nocodo. It is not ready for actual use and I first want to use it to build itself (well the core of it to build the rest of the parts).

replies(1): >>45034988 #
wjsekfghks ◴[] No.45034988[source]
To clarify, most of the execution (writing code or researching) is happening on the cloud. And we use Firestore as DB to store tasks. The app (both desktop and mobile) is just interface to talk to those backends. We are currently working to see if we can bring majority of the execution to local. Hope this makes it a bit clearer.
replies(1): >>45035027 #
brainless ◴[] No.45035027[source]
Thanks for the clarification.

Does this mean that my codebase gets cloned somewhere? Is it your compute or mine, with my cloud provider API keys?

replies(1): >>45035091 #
wjsekfghks ◴[] No.45035091[source]
If you use the app as is, it will be cloned to our server. If you choose to host your own server, it will be on yours
replies(1): >>45035589 #
1. brainless ◴[] No.45035589{3}[source]
OK thanks.