Armenia Tech Landscape

A curated directory of Armenian tech companies — from Yerevan-based startups to globally successful enterprises founded by Armenians.

🌐 Live Site

Visit the interactive catalog: Armenia Tech Landscape

📋 About

This project maps the Armenian tech ecosystem, covering companies across sectors like:

• AI & Machine Learning
• Cybersecurity
• Developer Tools & Cloud
• Fintech
• SaaS & Enterprise Software
• HealthTech, EdTech, and more

Each company is tagged by:

• Origin: Local (based in Armenia) or Global (founded by Armenians abroad)
• Size: Employee count ranges
• Type: Product or Service company
• Verticals: Tech sector tags

🏢 Companies

See the full catalog in the live site or browse contents/ for the raw YAML data.

🛠️ Contributing

Adding a Company

1. Create a YAML file in contents/ named after the company's website URL:

   • https://example.com → example.com.yaml
   • Run make fix-names to auto-correct filenames

2. Use this schema:

   name: "Company Name"
   origin: "local"          # or "global"
   employees: "11-50"       # 1-10 | 11-50 | 51-100 | 101-250 | 251-500 | +500
   company_type: "product"  # or "service"
   website:
     url: "https://example.com"
   links:
     - url: "https://linkedin.com/company/example"
       type: "linkedin"
     - url: "https://example.com/careers"
       type: "careers"
   description: "What the company does..."
   tags:
     - saas
     - ai-ml

   Link types: linkedin, careers, twitter, github, facebook, instagram, youtube, other

3. Tags must be from the allowed list in src/awesome_media/config.py. Add new tags there first if needed.

4. Run make build to verify and preview locally.

Development Setup

# Install Python dependencies
uv sync

# Build everything
make build

# Serve locally
make serve

# Frontend only
cd frontend
pnpm install
pnpm run dev

🤖 Automation Scripts

All scripts run via uv run python scripts/<script>.py.

Scrape from Trello board HTML

uv run python scripts/armenian_tech_scraper.py

Reads scripts/board.html (a saved Trello board export) and generates YAML files:

• Complete cards (with origin / employees / company_type labels) → contents/
• Incomplete cards → drafts/

Find incomplete YAMLs

# Console report
uv run python scripts/find_incomplete_yamls.py

# JSON report
uv run python scripts/find_incomplete_yamls.py --json report.json

# Only drafts
uv run python scripts/find_incomplete_yamls.py --only-drafts

Shows which fields are missing per file and distinguishes Checked vs Unchecked.

Auto-complete with Gemini (+ Google Search)

# 1. Set your key
cp .env.example .env
# edit .env and add GEMINI_API_KEY

# 2. Run (batches 20 companies per API call by default)
uv run python scripts/complete_yamls_with_gemini.py

# Process only 50, with 3s delay between batches
uv run python scripts/complete_yamls_with_gemini.py --limit 50 --delay 3

# Preview prompts without calling API
uv run python scripts/complete_yamls_with_gemini.py --preview --limit 30

# Save prompts to JSON for manual use
uv run python scripts/complete_yamls_with_gemini.py --save-prompts prompts.json

# Re-check files that were already processed
uv run python scripts/complete_yamls_with_gemini.py --recheck

Features:

• Batching (default 20 per call)
• Google Search grounding
• Model set via GEMINI_MODEL env var (default: gemini-3.1-flash-lite)
• Auto-retry with exponential backoff on rate limits
• Resumable via .gemini_progress.json
• Skips already-checked files unless --recheck is passed

Apply manual LLM output

If you ran the prompts elsewhere and have the JSON responses:

uv run python scripts/apply_llm_output.py --input responses.json

# Dry run first
uv run python scripts/apply_llm_output.py --input responses.json --dry-run

# Also improve existing contents/ files
uv run python scripts/apply_llm_output.py --input responses.json --include-contents

Accepts a JSON array (or object with results/companies/data/items key) where each element has an index matching the file order.

Fix filenames

make fix-names

Renames YAML files to match their website.url using the url_to_filename() convention.

📁 Project Structure

• contents/ — Company data (one YAML file per company)
• drafts/ — Incomplete YAMLs waiting for manual or LLM completion
• src/awesome_media/ — Python backend for validation and export
• frontend/ — Svelte SPA for the interactive web UI
• output/ — Generated site (HTML, JSON, assets)
• scripts/ — Automation tools (scraper, LLM completion, etc.)

📄 License

MIT