Async

What do they actually do

Async is a developer tool that runs coding agents in cloud containers to make code changes and deliver them as pull requests for you to review and merge. It combines a lightweight task tracker with automated code edits, build/lint/test runs, and for frontend work, preview deployments and screenshots so you can validate changes without running the repo locally (site, HN launch thread).

A typical flow is: you create a task, Async asks clarifying questions when needed, spins up an isolated environment, writes and tests the changes, then opens a PR with artifacts. You review and comment in‑app; the agent iterates based on your feedback. The product emphasizes parallel, “fire‑and‑forget” tasks you can manage from desktop or mobile, with GitHub PR integration (site, HN).

Async positions itself as an “AI‑native Linear”: not just tracking tickets, but executing them. The YC profile says it is open‑source and self‑hostable, aimed at removing bug‑fix/KTLO busywork for experienced teams who maintain mature codebases (YC profile).

Who are their target customer(s)

• Experienced IC at a startup/scaleup: Spends hours on small bug fixes and UI tweaks; context switching and local environment setup for tiny tasks interrupts feature work. Wants PR-ready changes without babysitting runs (site, YC).
• Engineering manager owning a papercuts backlog: Backlog of migrations, linting, and minor bugs drains team capacity. Needs to parallelize low‑value work without adding senior headcount (YC).
• Frontend team shipping frequent UI changes: Manual preview builds and slow QA for small UI fixes. Needs previews and screenshots to review changes without running the repo locally (site).
• Platform/DevOps owner responsible for CI/QA: Worried about breaking changes from automation. Needs isolated, auditable runs with full build/test verification before merge (site, HN).
• Security‑conscious teams or enterprises: Cannot send source code to third‑party services. Needs open‑source and self‑hosted options to keep code and agent execution on‑prem (YC).

How would they acquire their first 10, 50, and 100 customers

• First 10: Founder‑led pilots with YC/HN teams holding real papercut backlogs; onboard their repo, complete 2–5 tasks end‑to‑end (task → containerized edit → PR + preview) to prove value quickly (site, YC, HN).
• First 50: Publish concrete case studies (actual diffs, tests, preview artifacts) and offer a low‑friction GitHub App trial with templated tasks; seed developer channels and provide limited credits to convert interest into repo integrations (site, HN).
• First 100: Package a self‑hosted/open‑source agent and a verification playbook (full build/test runs, previews, audit logs). Sell short paid pilots to platform owners, integrate with GitHub/CI, and close with support SLAs for team‑wide rollout (YC, site).

What is the rough total addressable market

Top-down context:

There are roughly 27 million software developers worldwide, indicating a large spend pool for developer tools that improve velocity and automate maintenance (Evans Data, 2024).

Bottom-up calculation:

If Async targets ~1 million developers in teams that regularly handle bug‑fix/KTLO/UI tasks and charges about $80 per active user per month, annual TAM ≈ 1,000,000 × $80 × 12 = ~$960M.

Assumptions:

• Approximately ~4% of global developers fit the near‑term target profile (mature codebases, frequent small changes).
• Average paid usage is one active seat per developer at $80/user/month.
• Focus is on teams using GitHub‑style PR workflows where agent‑run build/test/preview adds value.

Who are some of their notable competitors

• GitHub Copilot (incl. Copilot for Pull Requests/Agents): IDE- and GitHub‑centric AI for code suggestions, PR summaries/reviews, and some agentic tasks that can open or update PRs. Strong editor/GitHub integration but not an out‑of‑the‑box issue→container→preview pipeline (Copilot PRs, docs).
• Cursor: AI‑first editor with cloud agents that can run code remotely and open GitHub PRs. Closer to Async on async cloud agents, but framed as an IDE + agent layer rather than a unified task→PR + preview workflow (Cloud Agents, GitHub integration).
• Sourcegraph (Cody/Amp): Repo‑aware assistants and automation with strong code indexing for large codebases. Emphasizes enterprise code search, controls, and automated multi‑repo changes over an integrated issue→container→PR preview loop (home, Cody/Amp blog).
• Replit (Ghostwriter/Agents): Cloud IDE with AI assistants/agents that edit, run, and deploy inside Replit workspaces. Optimized for in‑browser development and quick app shipping rather than automating PRs against existing external repos (Ghostwriter intro, agent integrations).
• CodeRabbit (PR automation): Automated PR review, summaries, and inline suggestions. Competes on reducing reviewer workload but typically doesn’t run full containerized builds or produce preview deployments (product).