Aperture

Aperture is a low-cost B2B agency lead pipeline for selling practical AI workflow implementation to international agencies.

The first motion is simple:

1. Build a daily list of B2B agencies from seed URLs, optional public search, and optional CSV imports.
2. Score them for likely AI workflow ROI.
3. Generate a manual approval batch for founder-led LinkedIn/email outreach.
4. Use OpenClaw only for capped enrichment on the best leads.

Stack

• FastAPI
• PostgreSQL
• Redis + Dramatiq
• OpenClaw for optional enrichment and drafting
• Amazon SES when production email sending is enabled
• Twilio WhatsApp only after explicit compliance checks

Prospecting Quick Start

Run from the repo root:

python ops\prospecting\discover_agencies.py --dry-run
python ops\prospecting\discover_agencies.py --source seed --seed-file ops\prospecting\agency_seed_urls.example.txt --max-results 50 --min-score 45
python ops\prospecting\build_agency_pipeline.py --dry-run
$today = Get-Date -Format yyyy-MM-dd
python ops\prospecting\build_agency_pipeline.py --no-search --input-csv "data\prospects\agency_research_queue_$today.csv" --max-sites 30

Generated prospect files are written to ignored data/prospects/.

API-backed discovery is optional. Configure Brave, Serper, Tavily, Exa, SerpAPI, or Google Programmable Search keys in .env when you want automated web-search fanout.

OpenClaw enrichment is off by default. Use it only for the best deterministic leads:

$today = Get-Date -Format yyyy-MM-dd
python ops\prospecting\build_agency_pipeline.py --no-search --input-csv "data\prospects\agency_research_queue_$today.csv" --max-sites 30 --openclaw-top-n 5

The operating runbook is in ops/prospecting/agency-lead-pipeline.md.

App Setup

1. Copy ops/.env.example to .env.
2. Update only the provider credentials you actually have.
3. Start the local services with Docker Compose.
4. Run openclaw onboarding only if you want optional AI enrichment.
5. Start the API and worker processes when using the backend.

Detailed setup is in docs/setup.md.