ai content engine

fully autonomous content pipeline - analyzes engagement, generates posts, schedules across 7 platforms. gets smarter every night.

get started

pick your agent framework:

git clone https://github.com/joshchoi4881/dropspace-agents && cd dropspace-agents && npm install

node setup.js --template dropspace-content-engine

node scripts/test-pipeline.js --app myapp

node scripts/run-self-improve-all.js --app myapp  # analyze + generate
node scripts/schedule-day.js --app myapp              # schedule for today

platforms

about

the core dropspace automation pipeline. every night, an LLM agent pulls analytics for all recent posts, scans X/Twitter for trending hooks in your niche, analyzes what's working vs what isn't, generates new posts with complete content (slideshows, text, captions), writes strategy notes that persist between runs, and schedules everything via dropspace. the feedback loop compounds - each cycle produces better content because it learns from real engagement data. supports visual (story-slideshow for TikTok/Instagram), text (text-single for Twitter, text-post for LinkedIn/Reddit/Facebook), and video (ugc-reaction and ugc-talking for TikTok/Instagram). the LLM owns all strategic decisions: which formats to use, what hooks to try, when to kill underperforming experiments.

features

• self-improving analytics loop
• content formats: story-slideshow, ugc-reaction, ugc-talking, text-post, text-single
• X/Twitter research integration
• correlation engine (auto-detects engagement patterns)
• format experiments with activate/kill/graduate lifecycle
• cross-platform strategy notes
• pluggable image providers (fal.ai, replicate, openai)
• multi-app support (one engine, many products)

requirements

• •dropspace account + API key
• •openclaw agent runtime
• •anthropic API key
• •image generation API key (fal.ai, replicate, or openai)
• •node.js 18+

content formats