Find potential customers from GitHub stargazer activity. Monitor any repositories for new stars, scrape historic stargazers on first run, and use AI to qualify leads automatically.

1. Historic Scraping: On first run, scrapes all existing stargazers from configured repositories
2. Ongoing Monitoring: A cron job monitors for new stars on your configured repositories
3. AI Lead Qualification: An AI agent researches each stargazer and scores them as a lead
4. Dashboard & Alerts: Qualified leads appear in your dashboard with periodic digest emails

1. Click Remix
2. Add environment variables:
   • OPENAI_API_KEY — for AI lead qualification
   • GITHUB_TOKEN — for accessing GitHub API (create one here)
3. Configure your repositories:
   • Edit GITHUB_REPOS array in github.cron.ts
   • Add repositories in "owner/repo" format (e.g., "facebook/react")
4. Configure email settings:
   • Edit RECIPIENTS in digest.ts
   • Customize PROMPT.txt with your ICP criteria
5. Open main.ts to view your dashboard

Note: The first run will scrape ALL historic stargazers, which may take time for popular repositories.

Repository Monitoring (github.cron.ts)

• Small Repositories (≤1,000 stars): Scrapes all stargazers in one run
• Medium Repositories (1,001-5,000 stars): Scrapes in chunks of 500 stars per run
• Large Repositories (5,001-10,000 stars): Gradual scraping with 200 stars per run
• Very Large Repositories (>10,000 stars): Skips historic scraping, monitors new stars only
• Ongoing Monitoring: After historic scraping completes, monitors for new stars since last check
• Tracks repository state and progress to resume scraping across multiple runs

Gradual Scraping Process

1. Assessment: Checks repository size and determines scraping strategy
2. Chunked Processing: Large repositories are processed over multiple cron runs
3. Progress Tracking: Maintains state between runs to resume where it left off
4. Automatic Completion: Transitions to new-star monitoring when historic scraping finishes

AI Agent (agent.ts)

• Researches each GitHub user's profile, repos, and linked sites
• Uses web search to learn about their company and role
• Scores them against your ICP defined in PROMPT.txt
• Returns {name, match, score, leadTypes, reasoning}

Storage (db.ts)

• Leads Table: Every lead is stored in SQLite with columns:
  • id — auto-incremented
  • timestamp — when first seen
  • input_data — the GitHub stargazer event(s) that triggered it
  • output_data — AI result
• Repository State Table: Tracks scraping progress per repository:
  • repo_name — repository identifier
  • last_checked — timestamp of last check
  • historic_scrape_complete — whether initial scrape is done
  • historic_scrape_progress — current page for gradual scraping
  • total_stars — total stars when scraping started

Dashboard (main.ts)

• Qualified leads (match=true) appear at the top
• Shows score and lead type tags (customer/hire)
• Click a lead to see full details and GitHub activity

Email Digest (digest.ts)

• Sends daily emails (1pm UTC) with new qualified leads
• Edit RECIPIENTS array to configure who receives them

• Repositories: Edit the GITHUB_REPOS array in github.cron.ts to add/remove repositories
• Lead Criteria: Edit PROMPT.txt to define your ideal customer profile
• Monitoring Frequency: Adjust the cron schedule in github.cron.ts (default: hourly)
• Email Schedule: Adjust the email digest schedule in digest.ts (default: daily at 1pm UTC)

Add repositories to monitor in the GITHUB_REPOS array:

export const GITHUB_REPOS = [
  "your-org/your-repo",
  "facebook/react", 
  "microsoft/vscode",
  "octocat/Hello-World", // Good for testing
  // Add more repositories here
];

• Gradual Scraping: Large repositories are processed over multiple cron runs to avoid timeouts
• Rate Limits: The system includes delays and batching to handle GitHub API limits
• Incremental Updates: After historic scraping, only new stars are processed
• Error Handling: Individual lead processing failures won't stop the entire job
• Resume Capability: System automatically resumes scraping after interruptions

• Small Repos (≤1,000 stars): Complete historic scraping in one run
• Medium Repos (1,001-5,000 stars): Scrape 500 stars per cron run
• Large Repos (5,001-10,000 stars): Scrape 200 stars per cron run (gradual)
• Very Large Repos (>10,000 stars): Skip historic scraping, monitor new stars only
• Batch Processing: Leads are processed in batches of 5 to ensure reliability
• Progress Tracking: System resumes scraping where it left off across multiple runs

Check scraping progress at /status:

• View status of all configured repositories
• See progress percentage for ongoing historic scrapes
• Monitor completion status and last check times

Use the test bar in the dashboard to evaluate any GitHub username instantly.

• "GITHUB_TOKEN is required": Add your GitHub token to environment variables
• Timeouts: Use smaller repositories or the system will automatically skip historic scraping
• Rate Limits: The system includes delays and batching to handle GitHub API limits
• Repository Not Found: Ensure repository names are in "owner/repo" format and are public