Open source · In production · Shipping real apps

The software company that lives in your repo

LARPCO is an open-source autonomous agent-company framework that runs inside GitHub Actions. It spawns a full cast of named AI agents who plan work, debate approaches, seek approvals, and ship code.

Corporate cosplay that actually ships. The PRs still merge.

Meet the team See how it works
Dottie Orchestrator
Xanthe PM Agent
Skye Design Agent
Blaze Eng Lead
Alma Loop Reviewer
Bindi Engineer Worker
Velma QA Reviewer
Pixie Feature Scout
Dot Learning Agent
Dottie Orchestrator
Xanthe PM Agent
Skye Design Agent
Blaze Eng Lead
Alma Loop Reviewer
Bindi Engineer Worker
Velma QA Reviewer
Pixie Feature Scout
Dot Learning Agent
What it is

Enterprise theatre.
Real engineering.

LARPCO spawns a full cast of AI agents inside your GitHub repo. They have roles, titles, and a chain of command. Under the costume, the architecture is solid. The bit just makes it more fun to run.

🏢

A company in your repo

Orchestrators, PMs, engineers, reviewers, scouts. They plan sprints, write PRDs, debate approaches, open PRs, and post standups to Slack. You're the CEO. That's your entire job.

🔁

Ralph Loops

Every engineer agent runs in a fresh context window with exactly one task. One file. Ships it. Goes home. The agent that writes code never reviews it. Separation of concerns, taken literally.

🧠

Compounding memory

Every completed loop writes its learnings back to a shared ruleset. Bugs get documented so they can't recur. The system gets smarter with every task it ships. Dot remembers everything.

💬

Slack-first human loop

Approval checkpoints via Slack. You approve scope, not implementation. A reaction emoji kicks off the pipeline. You get a ping when something needs a human. That's it.

🚦

Pre-execution plan review

Alma catches bad plans before any engineer spends tokens on them. A bad loop plan wastes every token downstream. Plans are checked before workers run. Always.

GitHub Actions runtime

No proprietary platform. No vendor lock-in. The pipeline runs on standard CI infrastructure. One script. Auditable. Forkable. Already works.

The cast

Your company.
Named and staffed.

Every LARPCO instance ships with a named company of agents. They have roles. They have vibes. They don't have personalities — they have a chain of command.

Name Role Vibe
Dottie Orchestrator Runs the whole show. Has seen everything.
Xanthe PM Agent Very into process. Writes long PRDs nobody asked for.
Skye Design Agent Sends mood boards. Cc's everyone.
Blaze Eng Lead Decomposes your tasks. Very confident about it.
Alma Loop Reviewer Catches everything. Trusts no one.
Bindi Engineer Worker One file. Ships it. Goes home.
Velma QA Reviewer Finds the bug. Every time. Not sorry.
Pixie Feature Scout Fires after every deploy with unsolicited ideas.
Dot Learning Agent Quietly updates the memory. Never speaks otherwise.

The full pipeline

A task enters as a Slack approval and exits as a merged PR with passing tests. You approve the what. The company handles the how.

Trigger
You approve a task on Slack
A reaction emoji. That's it. Dottie reads the approval and starts routing work.
Upstream
Xanthe — PM
Translates a goal into a PRD with acceptance criteria. Scoped to prose. No code.
Upstream
Skye — Design
Produces a component-level design spec from the PRD. Cc's everyone.
Planning
Blaze — Eng Lead
Reads the PRD + design spec. Produces atomic task briefs for each worker. Very confident about it.
Gate
Alma — Loop Reviewer
A separate instance. Never the Eng Lead. Reviews the plan for ambiguity, missing criteria, and architectural risk. Trusts no one.
Implementation
Bindi — Workers
Fresh context. One file. Opens a PR. Writes tests. Goes home.
Review Gate
Velma — QA
Never the engineer who wrote the code. Finds the bug. Every time. Not sorry.
Output
Shipped feature · Passing tests · Standup on Slack
You get a summary of what shipped. Pixie proposes what to build next. Dot quietly updates the memory.
Dottie never writes code. She reads state, routes to the right agent at the right time, and keeps your context load minimal. All orchestration logic lives in a single run-pipeline.mjs script. No proprietary platforms. The whole system runs on GitHub Actions and can be inspected, forked, and extended.
Positioning

The honest comparison

GSD says "no enterprise theatre." LARPCO says "we exclusively do enterprise theatre." The PRs still merge.

GSD LARPCO
Vibe Minimalist, efficient Corporate cosplay, chaotic good
Structure Commands you run Company that runs itself
Human role Developer CEO
Agents Subagents, anonymous Named cast, roles, chain of command
State .planning/ directory ORCHESTRATOR_STATE.md + git
Approvals Interactive prompts Slack reactions
Tone "No bs" "The bs is the feature"
Works? Yes Also yes
The bit is real

Under the costume

Every decision in the system was made to prevent a specific failure mode. The comedy is in the framing, not the output.

01

Fresh context, every task

Agents accumulate bias and drift when they carry context across unrelated tasks. Every worker spawns with a clean window and exactly the context it needs.

02

Implementer ≠ reviewer

An agent cannot objectively review its own output. Velma and Alma are always separate instances from the engineer. Non-negotiable.

03

Specialization over generalism

Xanthe writes better PRDs when she's never touched code. Blaze plans better when he doesn't also have to implement. Specialization compounds.

04

Plans gate execution

A bad loop plan wastes every token downstream. Alma exists to catch under-specified plans before any engineer spends tokens on them.

05

Learnings write back to memory

Dot updates the shared ruleset after every completed loop. Bugs discovered in production get documented so they cannot recur. The system improves with every task.

06

Humans approve scope, not implementation

You decide what to build and approve task scope. The agents decide how to build it. This keeps you strategic and the pipeline fast.

Early stage

We're early —
and that's the point

LARPCO is in active development, battle-tested on real products, and not yet available to external teams. We're building in public, learning fast, and looking for early partners who want to shape what this becomes.

b@punkyin.com →