Skip to content

73anthonyL/reconnaissance

Repository files navigation

Reconnaissance

Know your stealth competitors 3 months before they announce.

pip install -e . && recon scan --demo

What it does

Reconnaissance monitors GitHub commit patterns, developer community discussions, and hiring velocity to surface competitive threats before a company goes public with their direction. By the time a competitor announces, it's too late.


Example Output

──────────────────── RECONNAISSANCE BRIEFING ─────────────────────
  Idea: AI-powered competitive intelligence tool for founders
  Date: 2026-04-17

╭──────── ⏱  Window of Opportunity ─────────╮
│ 1 months  ● CLOSING FAST                  │
│ 1 CRITICAL  15 ELEVATED signals detected  │
╰────────────────────────────────────────────╯

Top Stealth Competitors

  1. Meridian AI  (score: 180 | sources: JOBS)
     → JOBS: Meridian AI — Founding ML Engineer
     → JOBS: Meridian AI — Senior Backend Engineer
     → JOBS: Meridian AI — Head of Growth

  2. Parallax Labs  (score: 153 | sources: JOBS)
     → JOBS: Parallax Labs — Product Designer
     → JOBS: Parallax Labs — Enterprise Sales Lead
     → JOBS: Parallax Labs — Senior Data Engineer

  3. Helios Analytics  (score: 142 | sources: JOBS)
     → JOBS: Helios Analytics — Lead ML Engineer
     → JOBS: Helios Analytics — VP of Engineering

╭────────────┬────────┬────────────────────────────────┬────────┬──────────────╮
│ Tier       │ Src    │ Title                          │  Score │ Flags        │
├────────────┼────────┼────────────────────────────────┼────────┼──────────────┤
│ CRITICAL   │ JOBS   │ Nexus Intelligence — ML Resea… │     70 │ NEW          │
│ ELEVATED   │ JOBS   │ Cascade Systems — AI Product … │     64 │ NEW          │
│ ELEVATED   │ JOBS   │ Nexus Intelligence — Founding… │     64 │ NEW          │
│ ELEVATED   │ JOBS   │ Meridian AI — Founding ML Eng… │     63 │ NEW          │
│ ELEVATED   │ JOBS   │ Meridian AI — Senior Backend … │     59 │ NEW          │
│ ELEVATED   │ HN     │ Ask HN: Is there a tool that … │     55 │ NEW          │
│ MONITOR    │ HN     │ Competitive intelligence for … │     38 │ —            │
╰────────────┴────────┴────────────────────────────────┴────────┴──────────────╯

Strategic Recommendations

  ■ CRITICAL
     Move to launch NOW. Compress your timeline by cutting scope —
     a shipped MVP beats a perfect product 3 months from now.

  ▲ ELEVATED
     Increase scan frequency to weekly. Identify the 1-2 features
     that would make switching costs prohibitive for your customers.

──────────────────────────────────────────────────────────────────

Install

git clone <repo>
cd reconnaissance
pip install -e .
recon scan --demo

Model (~90MB) downloads automatically on first run. No API keys required.


Usage

recon scan --demo                               # see output instantly, no setup
recon scan --idea "your startup idea" --save    # scan + save briefing to disk
recon scan                                      # interactive prompt
recon briefing --date 2026-04-17                # load a saved briefing

How It Works

  1. Ingest — three async scrapers query HN Algolia, GitHub Search, and a job posting feed in parallel
  2. Scoreall-MiniLM-L6-v2 embeds each signal; composite score = 50% semantic similarity + 30% momentum + 20% recency
  3. Diff — signals are compared against ~/.recon/state.json to surface only NEW and ESCALATING threats
  4. Brief — top 10 signals rendered as a Rich terminal table + optional Markdown file in ./briefings/

Architecture decisions (source selection, weight rationale, schema design): ARCHITECTURE.md
Non-obvious implementation choices: DECISIONS.md


Data Sources

Source Signal Momentum metric
Hacker News (Algolia API) Early founder discussions, Show HN posts Comment velocity
GitHub Search API Repos < 90 days old matching idea keywords Stars/day since creation
Job postings Hiring patterns before announcements Jobs/week per company

No API keys. No paid services. Fully offline after first model download.

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors

Languages