Agency Radar

Agency Radar is a low-capital product bet: a recurring procurement-intelligence feed built on open federal spending data.

The core idea is simple:

• pull a narrow market slice from USAspending
• summarize the highest-value agencies and recipients
• flag contracts expiring soon
• package the report into HTML, Markdown, and CSV
• sell the recurring intelligence to niche buyers

Why This Product

This is closer to a money machine than a generic app because:

• the data source is public
• the pipeline can be automated
• the output can be sold as a recurring report
• the product can start with one niche and one buyer profile
• it does not require a marketplace approval before it can be useful

It is not fully passive from day one. Nothing from zero is. But it can become low-touch much faster than a service business.

Current MVP

• config-driven buyer niches in config/profiles.json
• live award pulls from the official USAspending API
• summary analysis for top agencies, top recipients, expiring awards, and top recompete candidates
• change tracking for newly observed awards compared with the previous run
• static report generation to:
  • HTML
  • Markdown
  • CSV
  • JSON summary
• ranked buyer-target exports from live award data
• batch outreach draft generation for the highest-priority prospects
• publishable niche catalog in docs/catalog.html
• buyer-specific outreach generation from live report data
• scheduled automation template in .github/workflows/build-report.yml
• landing page in docs/index.html

Quick Start

List profiles:

PYTHONPATH=src python3 -m agency_radar.cli list-profiles

Build a report:

PYTHONPATH=src python3 -m agency_radar.cli build --profile federal_it

Outputs land in reports/.

Build the full public site catalog:

PYTHONPATH=src python3 -m agency_radar.cli build-site

That refreshes:

• docs/catalog.html
• docs/reports/
• docs/data/catalog.json

Generate ranked buyer targets:

PYTHONPATH=src python3 -m agency_radar.cli prospects --profile federal_it --top 12

Generate a batch of outreach drafts:

PYTHONPATH=src python3 -m agency_radar.cli pitch-batch --profile federal_it --top 5 --sample-report-url https://agency-radar.example/sample-report.html

Generate a personalized outreach draft:

PYTHONPATH=src python3 -m agency_radar.cli pitch --profile federal_it --company "Acme Federal" --first-name Alex --sender-name Luke

Productization Path

The fastest real monetization path is:

1. Pick one niche profile.
2. Generate one polished report.
3. Sell a weekly or daily version to a tiny group of buyers.
4. Add:
   • Slack delivery
   • email delivery
   • customized watchlists
   • SAM.gov opportunity data later

Current Sales Assets

• pricing and positioning in docs/pricing.html
• public niche catalog in docs/catalog.html
• go-to-market notes in ops/go-to-market.md
• launch blockers and minimal identity requirements in ops/launch-blockers.md
• outreach templates in templates/
• generated prospect sheets in reports/
• public sample contacts in data/sample_contacts.csv

Real outreach lists should live under data/private/, which is gitignored.

Honest Constraint

This is a viable subscription engine, not magic. To get paid, we still need one of:

• outbound sales
• a niche audience
• a marketplace
• a referral loop

What this repo does is remove the heavy lifting from the product side so distribution becomes the only real problem left.