YouTube Content Pipeline
End-to-end automation across multiple channels. Monitors source feeds, downloads at best available quality, trims and formats clips, generates thumbnails from key frames, then schedules and publishes via the YouTube Data API.
- Runs on a macOS LaunchAgent cron schedule
- Handles multi-channel publishing with per-channel config
- Thumbnail extraction using ffmpeg frame selection
Python · yt-dlp · ffmpeg · YouTube Data API · LaunchAgent
View on GitHub ↗
Social Media Analytics Scraper
Cookie-injected headless browser automation that extracts commenter and engagement data from authenticated sessions - no API keys or third-party services required. Runs monthly, deduplicates incrementally, and syncs structured JSON to Google Drive by file ID.
- Extracts Chrome session cookies via browser_cookie3
- Playwright drives the authenticated browser context
- Incremental dedup - only new entries added each run
- Overwrites Drive files in place; no duplicate uploads
Python · Playwright · browser_cookie3 · Google Drive API
View on GitHub ↗
Subscription Access Bot
Self-hosted bot that automates gated access via payment processing. Handles new member onboarding, tracks subscription state, and manages access automatically - removing it on lapse or cancellation. Deployed to cloud with persistent state and zero manual intervention.
- Stripe webhooks drive all state transitions
- Cloud-deployed with persistent subscriber store
- Automated access grant and revocation on payment events
Python · Stripe API · Telegram Bot API · Railway
View on GitHub ↗
AI Video Clipping Tool
Analyses long-form video and automatically segments it into short-form clips using AI-assisted scene and speech detection. Handles output formatting and packaging for multiple platforms without manual review of each clip.
- OpenAI API used for transcript-based segment selection
- ffmpeg handles all cutting and re-encoding
- Playwright automates any browser-based upload steps
Python · OpenAI API · ffmpeg · Playwright
View on GitHub ↗
Recruiter Outreach Engine
Lead enrichment pipeline that combines web scraping with company email pattern inference and confidence scoring. Built to support a targeted job search - achieved 89% email completion across 200+ contacts spanning four countries.
- Infers email format from known verified addresses per domain
- Confidence score attached to each inferred address
- Output as structured spreadsheet ready for outreach tools
Python · web scraping · openpyxl · pattern inference
View on GitHub ↗