LaunchRepo.

Works with

Launch your SaaS on directories with Claude Code

Claude Code reads files and can operate a browser through MCP — the two things LaunchRepo asks of an agent. Clone the repository, paste one prompt, and it works the directory playbooks in a Chrome you can see.

Claude Code is the terminal agent most LaunchRepo buyers already have open. It reads AGENTS.md and the playbooks like any other files in a project, runs the Python scripts that create the workspace and start the dashboard, and — once the Chrome DevTools MCP server is configured — fills in directory forms in your own signed-in browser. This page is the setup for exactly that; the workflow itself is described on the how-it-works page.

What Claude Code needs

The repository, cloned locally
After purchase you accept a GitHub invitation and clone the private repository. Open that folder in Claude Code; it reads AGENTS.md, the runbook and the playbooks as ordinary Markdown.
Python 3.10+ on macOS or Linux
The workspace script, record.py and the local dashboard are Python standard library only. Nothing else is installed; Windows is not supported.
A browser integration via MCP
Claude Code talks to Chrome through an MCP server. The repository’s BROWSER-SETUP.md uses Chrome DevTools MCP; you add it to Claude Code’s MCP configuration for the project.
Standing permissions, once per product
What the agent may do — free registrations, final submissions, verification emails, badges — is recorded in authorizations.json during onboarding. Claude Code respects it on every resume; spending stays at zero.

Connecting Claude Code to your Chrome

The principle from BROWSER-SETUP.md: use the browser you can see and supervise. The tracker never controls Chrome; Claude Code needs its own browser integration, and it should keep one long-lived connection during a batch rather than reconnecting per action.

Preferred route: enable remote debugging in the Chrome you already use (Chrome 144 or newer, via the inspect page), register Chrome DevTools MCP with its auto-connect option in Claude Code’s MCP configuration, then trigger a harmless action and approve Chrome’s connection prompt yourself in the intended window. A running MCP process alone proves nothing — the visible tab has to move.

Fallback: launch a dedicated Chrome with its own user-data directory and a local debugging port, sign in to the needed accounts in that window, and point the MCP server at that endpoint instead of auto-connect. Never both modes at once, never a public port.

Before any submission, the smoke check: open a blank tab, navigate to a public directory homepage, read its heading, confirm which tab moved, then sign in to a target directory and check the account menu shows the right identity. A completed OAuth popup is not proof of a destination login.

The start prompt

This is the public start prompt from START.md. Paste it into Claude Code inside the cloned folder; it creates your private workspace next to the checkout (your products and credentials never live in the delivery repository), starts the local dashboard and answers with a 127.0.0.1 link.

START.md — paste as-is, in English
Read AGENTS.md, then follow "Start the dashboard for the owner". Create my private
workspace if I do not have one yet, start the dashboard, and give me the link.

The playbooks and instructions are English Markdown, which is what coding agents handle best. Your product copy can be in any language a directory accepts; the agent answers you in your language.

From there the dashboard leads onboarding: add the first product, fill product.json and identity.json, grant permissions, run the browser smoke check. The prompts to hand back to Claude Code for a batch, a resume or a review are in the dashboard as well.

What a run looks like

The illustration below is the one from the home page: a small batch, each platform ending in the state the tracker actually observed. Real statuses come from your submissions.csv; nothing here is a measurement.

What still needs you

LaunchRepo is deliberately not a background service. These moments stop the run and are recorded as human handoffs; the agent never works around them:

CAPTCHAs and bot checks
Claude Code stops, records prepared_needs_human and the dashboard lists the platform under “needs you”. You solve it in the same Chrome window; no stealth changes, no evasion.
2FA and verification codes
Codes arrive in the mailbox you signed in to in the connected browser. Forwarding a code by hand is a supported fallback and is recorded as a human handoff.
Paid upsells
If the only usable route costs money, the platform is recorded as deferred_paid. Claude Code never authorises a purchase or starts a card-backed trial; you decide later with the note in front of you.
Confirming “live”
A thank-you page is not a listing. Marking a platform live requires a public URL and your confirmation that you looked at it.

No benchmark, no ranking: we claim compatibility for Claude Code because it meets the two requirements above, not because we measured it against other agents.

Questions

Does Claude Code need a special plugin for LaunchRepo?

No. It needs an MCP browser integration — the repository documents Chrome DevTools MCP — and Python 3.10+. Everything else is Markdown, JSON and CSV it reads like any project.

Does it run while I am away?

No. Claude Code works in your session and in a browser you can see. CAPTCHAs, 2FA and paid decisions stop the run; you start the next batch when you want.

Which model or plan do I need?

We make no claims about models or plans; agent usage is billed by your provider and not part of the licence. Any Claude Code setup that reads files and drives the MCP browser meets the requirement.

What about Claude Code’s own permission prompts?

They stay yours to configure. LaunchRepo adds its own layer on top — authorizations.json — that scopes free registrations, submissions, mail actions and badges per product.

Can I use Claude Code on one product and Cursor on another?

Yes. The licence covers unlimited agents; the workspace files are the same whichever agent reads them, and the tracker does not care who wrote a record.

One payment, every product, any agent

The launch price is €99.98, once. It rises 20 % after every 20 sales; the pricing page shows the live price, what is in the repository and the licence in short.

See pricing