LaunchRepoLaunchRepo is delivered as a private GitHub repository, so “updates” are simply new commits and tags in that repository. Buyers pull them with git; nothing is installed and nothing phones home. Release notes for the repository itself ship as CHANGELOG.md inside it.
Turn a finished run into a launch card
- Create a local 1200 × 675 result card in two styles. Live, pending review, queued and scheduled outcomes stay separate; existing listings and drafts never count as new submissions.
- Preview the public product name, copy or download the PNG, and open an editable X draft. Attach the image and publish yourself. Nothing uploads automatically; demo exports are clearly marked.
- Optional reminders appear after a completed run, away from active work and recent problem reports. Dismiss them or switch them off in Settings. Back up your workspace, dry-run the updater and restart the dashboard after applying.
Affiliate program in the dashboard
- New AFFILIATE.md in the repository: 30 % of the net price for every sale through your link (currently about €31, it rises with the price steps), €20 off for the person you recommend, SEPA payout on the 1st of the month after 60 days, with links to activation, your affiliate dashboard and the terms.
- Dashboard: an "Earn 30 %" link in the sidebar, a small card on the overview and a popup that explains the program once. It never appears on the first start, in a demo, during an active run or next to a testimonial request; "Remind me later" waits a week and Settings has an "Affiliate reminders" switch. The dashboard only shows links and sends nothing.
- Agents never bring up, activate or share your affiliate link.
- Updating: back up your private workspace, then run scripts/update-workspace.py in dry-run mode before you authorize an apply.
Share a testimonial from the dashboard
- New "Share a testimonial" form in the dashboard (Support & feedback, plus a card on the overview and product page at the right moments). It publishes a GitHub issue in AleksDoesCode/LaunchRepo through your signed-in GitHub CLI, or saves a draft when gh is unavailable. You choose the name, optional X handle, product and whether your current results are added; nothing is published without the explicit consent checkbox.
- The dashboard asks only after real results: the first live listing, a finished run with counted submissions, and five live listings. It never asks within three days of a problem report you sent. "Later" waits for the next milestone and at least two weeks; "Don't ask again" stops it for good. Demo workspaces never publish.
- Agents mention the option once after a finished run and never write or send a testimonial for you.
Product facts show what is known
- The Product facts step in the Setup tab now lists the facts already saved in product.json (name, website, audience, verified features, pricing with its check date, free-forever status, earliest launch date) under "What LaunchRepo knows". Required fields that are still empty stay under "Still missing", and unknown required facts get their own "Still unknown" list. Before, only the gaps were visible, so a completed step looked empty.
- Updating: back up your private workspace, then run scripts/update-workspace.py in dry-run mode before you authorize an apply.
Route recheck complete, welcome guide and settings
- Route recheck part 2 is complete: all 349 catalog IDs carry a dated check against official public rules and 239 guides were rewritten. The catalog now has 299 active guides: 86 unconditional free routes (was 57), 157 conditional-free routes with visible prerequisites (was 77), 52 currently unavailable or blocked and 4 research-only. 24 more mandatory-paid routes became exclusion tombstones with free alternatives, 50 in total.
- Welcome guide in the dashboard: five short steps explain what LaunchRepo does and never does, the words you meet everywhere and your workspace checks, then create the first product from a human name with a live folder name, your preferred sign-in for directory accounts (Google first, or email and password) and a reminder that the agent needs the matching mailbox.
- New Settings page: theme, restart or continue the welcome guide, and the automatic problem-report choice moved there from Support & feedback.
- Updating: back up your private workspace, then run scripts/update-workspace.py in dry-run mode before you authorize an apply.
Neutral wording in dated catalog audits
- The dated September 14 catalog audits now use neutral wording and no longer name example products. Their platform findings are unchanged.
- Includes everything from 1.6.0: route recheck part 1, the new dashboard and opt-in support reports.
Route recheck part 1, new dashboard and support reports
- Rechecked 165 platform guides against official public rules. As of 2026-09-23 the catalog has 323 active guides: 57 unconditional free routes (up from 15), 77 conditional-free, 31 currently unavailable and 158 still research-only. 21 more mandatory-paid routes are excluded. The remaining guides follow in the next update.
- New local dashboard with guided setup, a status overview per product and run details.
- Support & feedback page: report problems, questions and feedback as GitHub issues, with a Report button on blocked submissions. Your agent can also report blockers automatically, but only if you turn that on; it is off by default.
Clearer run onboarding checklist
- The remaining dashboard checklist and prompt-card copy now match persistent runs. Existing saved checklist progress is preserved.
- Product, browser and permission checks come before planning; a first submission is not required to create a run.
Run goals, clearer browser setup and resumable onboarding
- Plan a target of 10 new confirmed submissions, with at most 2 workers and a local hardware recommendation. Separate browser control is required for parallel work; otherwise use one worker.
- Save the run and resume its remaining target. Copy the goal prompt to your agent; use /goal only when the host supports it. Confirmed new submissions count once per platform; drafts and existing listings do not.
- English onboarding, dashboard guidance and email now lead through Chrome DevTools MCP setup on Windows, macOS and Linux, including a conditional macOS Full Disk Access check. After saving a confirmed outcome, close only the assigned submission window.
- Native Windows and Linux browser connections still require the smoke check on the owner’s machine. Provider limits and human challenges can interrupt a run.
Windows compatibility, local dashboard and 347 playbooks
- The tracker and local dashboard now include native Windows compatibility with Python 3.10+, alongside macOS and Linux. Windows setup, file locking and UTF-8 file handling no longer depend on POSIX-only behavior.
- The release archive includes the local dashboard, updated onboarding and all 347 platform playbooks. PowerShell setup and Chrome connection instructions are included; a working browser integration is still required for submissions.
- Native Windows end-to-end validation is pending. The browser smoke check verifies the connection on your own machine before a submission.
Launch-date floor and representation briefing
- Owner-set earliest launch date in product.json; record.py requires --launch-date for “scheduled” and refuses anything earlier than that floor.
- ONBOARDING.md now asks when launches should go live and states plainly that free-tier directories can take several weeks to review and publish.
- New PR-BRIEFING.md: a short session that shows the owner factual representation examples (maker bio, launch intro, “why I built this”) for correction before any personal fact reaches a public listing.
- Approved examples and corrections are kept in templates/project/representation.md.
Local dashboard and one-prompt start
- scripts/dashboard.py and dashboard/: a zero-dependency local web UI (Python standard library, 127.0.0.1 only) showing products, what the agent knows about each one, onboarding checks, submission states grouped by needs-you / problems / in review / scheduled / live, follow-up dates, the catalog with playbooks, agent prompts and the docs.
- Product JSON and notes can be edited in the browser; outcomes are recorded through record.py with identical validation. --demo serves a throwaway sample workspace.
- START.md and prompts/start.md: one pasted prompt makes the agent create the private workspace, start the dashboard and hand back the link; AGENTS.md contains the exact procedure with verification.
- scripts/init-workspace.py copies dashboard/, DASHBOARD.md, START.md and PR-BRIEFING.md into new workspaces.
346 platform playbooks
- New platform playbooks, including CodeHype and additional launch boards, AI, developer-tool and startup directories — 346 in total.
- Catalog, platform index and search tool updated to include every new entry.
- To update an existing workspace: copy the new playbooks and the updated catalog and index files; keep workspace.json, projects/, credentials and submission records.
Priority evidence and catalog screening
- Dated, claim-specific priority evidence, current tracker overlays and score explanations; unstructured inputs stay readable but cannot earn unsupported SEO or audience credit.
- Full catalog HTTP screening snapshot with explicit unknowns plus a bounded listing-candidate screen — discovery success is not listing indexation.
- Public sample audit references added on the main branch; previously issued immutable release archives keep their original contents.
275 platform playbooks
- Platform playbook library expanded to 275 files with submission steps, required assets, account setup, badge placement, follow-up and troubleshooting.
- Every catalog entry now links to its own playbook; new platforms added.
- Platform selection simplified by product category and name.
- Product credentials and execution history stay in your separate private workspace; follow PRIVATE-WORKSPACE.md before any push.
First immutable release
- Canonical immutable GitHub release of the private customer repository (v1.0.0 was the initial non-immutable publication).
- 222 directory candidates in the catalog, 30 of them with detailed reviewed playbooks; the remaining entries are marked as research candidates, not verified free routes.
- Guided onboarding for product, browser, verification email, private credential storage and website badges; workspace and tracker tools (record.py, submissions.json).
- No pilot identities, credentials or execution history in the customer archive.
How updates reach you
After purchase you clone the repository once. New playbooks, catalog and index files arrive as normal commits — pull them whenever you like.
Your products, credentials and submission records live in your own private workspace next to the checkout, so an update never touches them. Keep workspace.json, projects/ and your tracker files; copy the new playbooks and catalog over.
Updates we publish are included for everyone who bought. We do not promise a specific update, cadence or future scope — see the terms of service.