ph_ai_tracker

A small Python library that finds trending AI products on Product Hunt.

• Primary: Product Hunt GraphQL API v2
• Secondary: Web scraping fallback (BeautifulSoup) for when the API is unavailable / rate-limited

Install (dev)

poetry install

Install (PyPI)

pip install ph-ai-tracker

Optional (faster HTML parsing with lxml):

pip install "ph-ai-tracker[lxml]"

Quickstart

from ph_ai_tracker import AIProductTracker

tracker = AIProductTracker(
    api_token="YOUR_PRODUCTHUNT_TOKEN",  # optional
    strategy="auto",                    # api | scraper | auto
)

result = tracker.get_products(search_term="AI", limit=20)
print(result.to_pretty_json())

Or from the terminal (prints pretty JSON):

poetry run python -m ph_ai_tracker --strategy scraper --search AI --limit 10

By default, each run is also persisted to SQLite at ./data/ph_ai_tracker.db.

poetry run python -m ph_ai_tracker --strategy scraper --db-path ./data/ph_ai_tracker.db

Use --no-persist to skip DB writes for one-off runs.

Scheduler (Sprint 2)

Run one scheduled cycle (scrape + persist) via the scheduler command:

poetry run ph-ai-tracker-runner --strategy scraper --search AI --limit 20 --db-path ./data/ph_ai_tracker.db

Environment variables supported by scheduler runtime:

• CRON_SCHEDULE (default: 0 */6 * * *)
• TZ (default: UTC)
• PH_AI_TRACKER_STRATEGY (default: scraper)
• PH_AI_TRACKER_SEARCH (default: AI)
• PH_AI_TRACKER_LIMIT (default: 20)
• PH_AI_DB_PATH (default: ./data/ph_ai_tracker.db)
• PRODUCTHUNT_TOKEN (required for API strategy)
• PH_AI_RETRY_ATTEMPTS (default: 2)
• PH_AI_RETRY_BACKOFF_SECONDS (default: 2)

Cron helper files are available in scripts/cron/.

Docker + Persistent Volume (Sprint 3)

Build and start the scheduler container:

docker compose up -d --build

Watch scheduler logs:

docker compose logs -f scheduler

The SQLite database is stored on the named Docker volume ph_ai_tracker_data at /data/ph_ai_tracker.db in the container.

Stop without deleting data:

docker compose down

If you remove containers and start again, data persists because the named volume is reused.

After installing from PyPI, you can also run:

ph-ai-tracker --strategy scraper --search AI --limit 10

Trending behavior

• --strategy api defaults to Product Hunt's RANKING order.
• The API client first attempts the artificial-intelligence topic; if the schema/topic query fails, it falls back to global posts and applies a client-side filter.

Product Hunt API token

Do not hardcode tokens in code or commit them to git.

Set your token as an environment variable:

export PRODUCTHUNT_TOKEN="<your_token>"
poetry run python -m ph_ai_tracker --strategy api --search AI --limit 10

Notes

• This project is intended to run well on PyPy 3.
• All tests are offline and use mocks/fixtures (no real network).
• Supported Python versions: 3.10–3.14
• Operational guide: see RUNBOOK.md
• Convenience commands: make test, make runner, make docker-up, make docker-logs
• Demo script: scripts/demo_pipeline.sh
• Requirements mapping: REQUIREMENTS_TRACEABILITY.md
• Final handoff checklist: SUBMISSION_CHECKLIST.md