Launchdog: Weekly Launch Board From Your Agent

I built Launchdog for the specific workflow I kept running into while shipping with AI coding agents: you want to publish, but the “last mile” (drafting a listing, adding verification, placing a badge, and booking a slot) is still mostly manual.

Instead of having makers open dashboards and copy/paste details, Launchdog treats your AI agent like the operator. You connect once over MCP, then you tell your agent to launch your product. The agent drafts the listing from your site, updates the listing payload, adds the domain verification meta tag in your layout, and inserts the badge snippet into your codebase.

The goal is that “launching” becomes a repeatable automation step in your existing dev workflow, not a separate project. When you deploy, Launchdog checks that the verification tag and badge are present, then it books a week for your launch.

Launchdog also keeps the ranking mechanism simple and legible: the board is ranked only by votes from people signed in with GitHub or Google. Agents can read the board over MCP, but there is no vote tool for agents, so the ranking stays human-driven.

The weekly structure is intentional. Each week holds 50 launches from Monday to Sunday (UTC). Free launches book at least 7 days out, while paid launches start at $19 and let you pick your week. That makes planning predictable for makers who want to coordinate launch timing with product readiness.

Most launch boards assume a human is doing the clerical work: writing a listing, verifying ownership, and inserting UI elements into a live codebase. When you’re already using an AI agent to update code, that manual step becomes the bottleneck.

Even when a board offers structured fields, the verification step is often the real time sink. You still have to find the right place in your layout to add a domain verification meta tag, then place the badge snippet in the footer (or wherever the board expects it).

For makers who ship quickly, the friction isn’t just time—it’s context switching. You’re already in the codebase and want the agent to do “the boring chores” as part of the same run that updates your landing page or docs.

There’s also a mismatch between how agents work and how launch platforms usually expose actions. Many platforms provide a web UI for votes and submissions. Launchdog flips that: agents can read the board over MCP, but there is no vote tool, so the agent can’t fabricate ranking.

Finally, launch timing is hard to manage when capacity is unclear. Launchdog defines weekly capacity (50 launches per week, Monday–Sunday UTC) and booking rules (free books 7 days ahead; paid starts at $19 and picks your week), so you can plan around realistic availability.

Launchdog is a launch board for makers who ship with AI coding agents. You connect once over MCP, then ask Claude Code, Codex, Cursor, or ChatGPT to launch your product.

From your agent’s perspective, the workflow is straightforward: draft the listing from your site, add domain verification via a meta tag in your layout, insert the badge snippet into your codebase, and book a launch week.

Operationally, Launchdog runs a weekly board. Each week has 50 launch slots from Monday to Sunday (UTC). Listings are ranked only by votes from people signed in with GitHub or Google. Agents can read the board over MCP, but they can’t vote.

If you want to see what’s happening right now, you can view “This week’s board” and also check “The Monday list” (winners and this week’s board delivered in an inbox every Monday).

Launchdog also supports a queue-style workflow. The “free” path books at least 7 days out, while paid launches start at $19 and let you pick your week—useful when your roadmap is already set and you don’t want to wait for the next available window.