Clarity

Company intelligence API for sales teams and AI agents. One API call returns structured intelligence about any company, including signals, contradictions, tech stack, hiring patterns, and a personalized outreach email.

What it does

Clarity takes a target company domain and your company domain, researches both in parallel, and returns:

• Company profile with industry, stage, and description
• Sales signals with implications (e.g., "just raised Series C" -> "budget available for new tooling")
• Contradiction detection between what the company says and what evidence shows
• Tech stack extracted from GitHub repos
• Hiring patterns and what they imply about priorities
• Bidirectional relevance scoring (is your product relevant to them? are they the right customer for you?)
• Suggested outreach email that references specific findings

Quick start

# Clone and set up
git clone https://github.com/deepgori/clarity.git
cd clarity
python -m venv venv
source venv/bin/activate
pip install -r requirements.txt

# Configure
cp .env.example .env
# Add your OpenAI API key to .env

# Run
python main.py
# Open http://localhost:8000

API

POST /api/company

Analyze a company and return structured intelligence.

curl -X POST http://localhost:8000/api/company \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -d '{
    "domain": "datadog.com",
    "seller_domain": "sentry.io",
    "context": "focus on their enterprise monitoring gaps"
  }'

Request fields:

Field	Required	Description
domain	Yes	Target company domain
seller_domain	No	Your company domain (Clarity auto-extracts what you sell)
context	No	Extra context for the analysis

Response:

{
  "success": true,
  "intelligence": {
    "company_name": "Datadog",
    "domain": "datadog.com",
    "what_they_do": "Cloud monitoring and security platform...",
    "industry": "Cloud Infrastructure / DevOps",
    "stage": "Public (NASDAQ: DDOG)",
    "signals": [...],
    "contradictions": [...],
    "tech_stack": ["Go", "Python", "TypeScript", ...],
    "hiring_signals": [...],
    "sales_strategy": {
      "recommended_angle": "...",
      "relevance_score": 0.85,
      "relevance_reasoning": "..."
    },
    "overall_confidence": 0.9
  },
  "suggested_email": "...",
  "processing_time_ms": 12000
}

POST /api/compare

Same as /api/company but also generates a generic email for side-by-side comparison.

Architecture

Request -> Parallel fetch (website + news + GitHub + seller) -> AI synthesis -> Response

Data sources:

• Website content via Jina Reader (with trafilatura fallback)
• News via DuckDuckGo (with NewsAPI fallback)
• GitHub repos, languages, and org data via GitHub API
• Seller website for bidirectional matching

AI layer:

• GPT-4o with structured JSON output for intelligence synthesis
• GPT-4o for personalized email generation
• Contradiction detection across data sources
• Bidirectional relevance scoring

Configuration

Variable	Required	Description
OPENAI_API_KEY	Yes	OpenAI API key
CLARITY_API_KEY	No	API key for authentication (if not set, auth is disabled)
CLARITY_GITHUB_TOKEN	No	GitHub token for higher rate limits
NEWS_API_KEY	No	NewsAPI key for news fallback

License

MIT