Most OSS projects I've seen die on day two. The tool is fine. The README is decent. The launch post got 40 upvotes. Then the author never posts again, because posting manually every day across Bluesky, Mastodon, Dev.to, and Hashnode is genuinely tedious work that compounds across multiple projects.
I built marketing-pipeline to handle the recurring part. After the initial setup, a GitHub Actions cron runs at 14:00 UTC on weekdays and rotates through your projects, angles, and channels — picking whichever angle was used least recently so you're not repeating yourself.
Onboarding a new project is one command:
marketing onboard --name my-tool --repo owner/repo --kind mcp-server
That fetches the README, sends it to Claude, and writes problem, facts, and content angles to projects.yml. The kind field matters: an mcp-server gets submitted to MCP Registry, Smithery, and Glama; a browser-extension routes to Chrome Web Store and Firefox AMO; a claude-skill goes to awesome-claude-code.
One thing I was firm about: the anti-slop gate in pipeline/antislop.py hard-rejects specific tokens before anything is published. "Excited", "game-changer", "unlock", "AI-powered", emoji, hashtags, exclamation points, rhetorical questions — any of those in a draft and it gets thrown out before posting. The channel length limits are enforced the same way: 280 chars for X, 300 for Bluesky, 500 for Mastodon, 150–400 words for Dev.to and Hashnode.
You need an Anthropic API key and credentials for at least Bluesky, Dev.to, and Hashnode. Mastodon and Slack are optional.
github.com/robertnowell/marketing-pipeline
Top comments (0)