$ man apollo-wiki/hacker-track

☾

Apollo API for sourcing. Supabase (Postgres) as your data warehouse. Claude Code for writing and iterating on scripts. A Mac Mini (or any always-on machine) running cron jobs. That is the entire stack. No SaaS platform fees, no per-seat pricing, no vendor lock-in. You own the data, the logic, and the infrastructure. Total monthly cost: Apollo free tier ($0) + Supabase free tier ($0) + Mac Mini electricity (~$5). The code lives in your repo. The data lives in your database. The pipeline runs while you sleep.

PATTERN

Never search Apollo for everything at once. Batch by persona. Run 1: search for VP/Director of Marketing at SaaS companies, 50-500 employees, US-based. Write results to Supabase. Run 2: same filters but VP/Director of Sales. Run 3: same but RevOps and Growth titles. Each batch gets its own sourcing run, its own title filter pass, and its own dedup check against existing contacts in Supabase. This prevents credit waste (no duplicate searches) and keeps your data organized by persona from the start.

PATTERN

Apollo rate limits are generous but real. 100 requests per minute on the free tier. Your enrichment script needs to respect this. Pattern: pull a batch of un-enriched contacts from Supabase (SELECT WHERE enriched_at IS NULL LIMIT 50). Run the enrichment calls with a 1-second delay between requests. Update Supabase with results. Log failures for retry. Run this on a 15-minute cron. You process 50 contacts every 15 minutes, 200 per hour, 4,800 per day. That is production throughput for most SMB and mid-market pipelines.

PRO TIP

You do not need to be a Python expert to build this pipeline. Claude Code writes the scripts. Describe what you want: "write a Python script that searches Apollo for VP of Marketing at SaaS companies with 50-500 employees, filters by title allowlist, deduplicates against existing Supabase contacts, and inserts new contacts." Claude Code generates the script, handles the API calls, writes the SQL, and even sets up the cron job. You review, test, and ship. The barrier to building this pipeline is not technical skill. It is knowing the architecture.