LaunchPad: One Dashboard for a 50-Channel Product Launch

The launch-day spreadsheet problem

Every indie product launch looks the same behind the scenes: a giant spreadsheet of channels — Reddit, Product Hunt, Hacker News, Indie Hackers, Twitter, LinkedIn, Discord, Telegram, newsletters — with columns for who posted what, where, when, and how it performed. It's manual, error-prone, and it falls apart exactly when things get busy: on launch day.

LaunchPad replaces the spreadsheet with a model. A launch is a Campaign; a Campaign generates a phased checklist from pre-seeded templates — 33 items for a SaaS launch, 29 for a mobile app, 28 for open source — split across pre-launch, launch day, post-launch, and ongoing. Every post, account, and engagement metric is a first-class database record.

AI writes it, Playwright posts it, Celery watches it

Content generation runs through OpenRouter: per-platform titles and bodies, with A/B variants, every generation tracked and attributable. A Reddit post and a LinkedIn post about the same feature come out sounding native to each platform, not copy-pasted.

Actual posting is handled by a Playwright browser pool — real browser automation, with Reddit as the deepest-built path, including subreddit targeting through a dedicated audience-fit model. Celery Beat schedules everything, and engagement snapshots build a time series per post, so 'how did the launch go' is a chart, not an archaeology project.

• ● Channel-agnostic engine — channels are data, not code — adding a platform extends a model, it doesn't fork the codebase.
• ● Self-hosted discovery — a bundled SearXNG instance powers research and monitoring without paid data APIs.
• ● Geographic intelligence — a GEORANK sub-module adds local and geographic ranking intelligence for location-sensitive launches.

Built the way we build everything

LaunchPad runs as an eight-service Docker stack — Django and DRF at the core, React with a component-driven UI on top, PostgreSQL and Redis underneath, Celery doing the heavy lifting. It's the same production architecture we ship for clients, applied to our own launch problem. We build tools we need first, harden them in real use, then productize.