SaaS Maker / Launchkit
Your agent launches the product. Free.
Paid launch kits sell a git repo of directory guides for around €100. This is the same shape — a product brief, a run planner, 330 per-platform submission playbooks, an evidence tracker, and a run prompt — built on the public launch catalog of 1163 destinations. No account, no payment, no upsell.
Quickstart
git clone --depth 1 https://github.com/sass-maker/saas-maker.git
cd saas-maker
cp tooling/launchkit/brief.template.md brief.md # fill once
node tooling/launchkit/scripts/plan-run.mjs --brief brief.md --n 10 --out run-plan.json
# paste tooling/launchkit/prompts/run.prompt.md into your agent
node tooling/launchkit/scripts/report.mjs --tracker tracker.json The five pieces
- Product brief
- Facts, approved claims, assets, and agent permissions — written once, reused every run.
- Run planner
- Picks ≤N suitable routes, free before conditional, DR-ranked. When fewer routes honestly fit, the target shrinks — it never pads with paid or unverified routes.
- Playbooks
- 330 per-platform guides: requirements, preparation, workflow, watch-fors, and what counts as live — every claim citing the public page it came from.
- Tracker
-
A saved draft is
prepared, a confirmed date isscheduled, a resolving public URL islive. States carry evidence or they don't count. - Run prompt
- Paste-ready instructions that make the agent stop for logins, CAPTCHAs, payments, and any permission the brief didn't grant.
For agents
Everything is machine-readable without cloning:
- /launchkit.json — the kit manifest
- /launchdesk.json — the full destination catalog
- /launchdesk-playbooks.json — playbook index;
each playbook also at
/launchdesk/<domain>.json - /llms.txt — site-level index
Honest grade: playbooks are researched from cited public pages, not observed submission runs. The schema says so on every file, and the tracker keeps your own run's evidence separate from research.