HomeCompaniesReplicas
Replicas

Background coding agents for engineering teams

Replicas is a background coding agent that allows you to delegate tasks to Claude Code or Codex from Slack, Linear, GitHub, and more. Each task works inside of a sandboxed virtual machine, allowing teams to heavily parallelize work and burn through tickets. Replicas agents output higher quality PRs via feedback loops - each task automatically is notified and fixes CI failures and reads all code reviews. With Environments, engineering teams can configure agents to run their entire codebase locally, verify its work, and provide preview URLs.
Active Founders
Connor Loi
Connor Loi
Founder
Building the best way to use background coding agents ;) waterloo cs, my all-time fav movie is kung fu panda 2
Company Launches
Replicas - End-to-End Background Coding Agents
See original launch post

Hi there, I'm Connor from Replicas :)

Recently, you might have seen Stripe and Ramp release blogs outlining their in-house background coding agents. This showed us that background agents, with the proper tooling, can be relied on to ship critical production code.

Replicas is a background coding agent that can run your codebase locally.

We’re used by 20+ YC startups at all team sizes.

uploaded image

The Problem

Background coding agents are a huge productivity boost for clearing tickets and shipping faster.

The existing solutions - CC/Codex on Cloud, Cursor Agents - lock you into one harness and give you no control over the environment.

This defeats the entire point of "background" agents. A powerful background agent should run everything locally, verify its work, and iterate until it's satisfied.

How Replicas Works

Replicas helps you delegate to as many background agents as you need:

- Bring your own Claude/Codex credentials

- Each agent gets a sandbox VM with your code cloned

- The agent installs deps, spins up databases, runs services, and verifies its work locally

- When it's done, it opens a PR for you to review

uploaded image

Replicas can be triggered from Slack, Linear, GitHub (PR reviews, CI/CD failures), the Dashboard, or API.

The biggest difference is Environments. You can configure the following:

- Files

- Env vars

- Start hooks

- Warm hooks

- Skills/MCPs

It's built so that its flexible to any kind of codebase, allowing Replicas to run everything locally just like a human engineer would.

uploaded image

Replicas can also create Previews:

uploaded image

Meaning you can check out services (e.g web app on 3000) running on the sandbox's VM from your computer with just a click :)

Check out this writeup for Replicas V1 for more details:

https://www.replicas.dev/notes/replicas-v1

Demo Video

https://youtube.com/watch?v=uyH77T-pcXc

What's Next

- Automations + API: Improving our API and adding an automations menu, e.g having Replicas run every night.

- Recursive: Replicas can break down super complex tasks by creating more Replicas instances.

- Lowering Latency: For larger teams, improving our warm pool and snapshotting infrastructure to make spinning up Replicas feel instant (more than it already does)

- Long-term Memory

- and what our customers ask for :p

What I'm asking from you

- Sign up (https://replicas.dev) or reach out (connor@replicas.dev)

- Check out [this writeup](https://www.replicas.dev/notes/replicas-v1) for Replicas V1

- Upvote if this resonates with you!

Jobs at Replicas
San Francisco, CA, US
$150K - $250K
2.00% - 4.00%
Any (new grads ok)
Replicas
Founded:2026
Batch:Spring 2026
Team Size:1
Status:
Active
Primary Partner:Harj Taggar