Skip to content

README.md

Latest commit

 

History

History
173 lines (120 loc) · 5.6 KB

README.md

File metadata and controls

173 lines (120 loc) · 5.6 KB
Raw

Prospect Analyzer

AI-powered prospect research agent that analyzes websites and generates branded HTML reports. Uses Anthropic's Managed Agents API to perform deep analysis with configurable industry-specific criteria.

What it does

  1. Takes a website URL as input
  2. An AI agent crawls the site, analyzes content, products, and compliance signals
  3. Produces a detailed Markdown analysis with a readiness score (0-10)
  4. Generates a branded HTML report suitable for sharing with prospects
  5. Optionally deploys reports to a static hosting service (Netlify)

Fully configurable — change industry_config.py to analyze any industry (compliance, security, UX, sustainability, etc.).

Prerequisites

  • Python 3.10+
  • Anthropic API key with Managed Agents access
  • (Optional) Netlify site for report hosting

Quick Start

  1. Clone and install dependencies

    git clone https://github.com/YOUR_USERNAME/prospect-analyzer.git
cd prospect-analyzer
pip install -r requirements.txt

  2. Set your API key

    cp .env.example .env
# Edit .env — add your ANTHROPIC_API_KEY

  3. Create the Managed Agent

    python setup_agent.py

    This provisions an agent and environment via the Anthropic API and saves the IDs to config.json.

  4. Customize for your use case

    Edit industry_config.py to define what the agent analyzes (see Configuration Examples below).

  5. Run an analysis

    python run_analysis.py https://www.example-shop.com

    Output is saved to output/{domain}-analysis.md.

  6. Generate an HTML report

    python run_analysis.py https://www.example-shop.com --html

    Or batch-convert all existing analyses:

    python generate_site.py

Deploy to Netlify (optional)

./deploy.sh

Copies HTML reports to your Netlify-connected repo and pushes.

Configuration

Environment variables

Variable Required Description
ANTHROPIC_API_KEY Yes Anthropic API key
COMPANY_NAME No Your company name for HTML reports
CONTACT_EMAIL No Contact email in report footer
COMPANY_WEBSITE No Website URL in report footer
CTA_URL No Calendly/booking link for CTA button
NETLIFY_BASE_URL No Base URL of your Netlify site
PROPOSALS_DIR No Path to your Netlify-connected git repo

config.json

Contains Managed Agent configuration (agent ID, environment ID, model). Created by setup_agent.py.

Project structure

prospect-analyzer/
├── setup_agent.py       # One-time agent creation via Anthropic API
├── run_analysis.py      # Run analysis on a target URL
├── html_generator.py    # Markdown -> branded HTML conversion
├── generate_site.py     # Batch HTML generation + index page
├── deploy.sh            # Deploy HTML reports to Netlify
├── config.json.example  # Agent config template
├── .env.example         # Environment variables template
└── requirements.txt     # Python dependencies

How the agent works

The Managed Agent uses Anthropic's hosted agent infrastructure with:

  • SSE streaming for real-time output
  • Tool use for web browsing, content extraction, and analysis
  • Structured output with readiness scoring across multiple dimensions

The agent evaluates the target website based on the criteria defined in industry_config.py. The default configuration performs a general readiness assessment, but you can customize the system prompt, scoring regex, and report branding for any industry.

Customizing for your industry

Edit industry_config.py to change:

  • SYSTEM_PROMPT — What the agent analyzes (compliance criteria, scoring rubric)
  • USER_MESSAGE_TEMPLATE — How the analysis request is phrased
  • REPORT_*_ — HTML report branding (title, subtitle, score label)
  • BRAND_REGEX / SCORE_REGEX — How to extract metadata from analysis output

Configuration Examples

Below are two examples showing how to adapt industry_config.py for different industries. You only need to override the key variables — everything else uses sensible defaults.

E-commerce UX Audit

Analyze online stores for usability issues and conversion optimization opportunities.

SYSTEM_PROMPT = """You are a UX and conversion optimization expert.
Evaluate the given e-commerce website across: navigation clarity,
product page completeness, checkout friction, mobile responsiveness,
page load indicators, and trust signals (reviews, security badges).
Score each dimension 1-10 and provide an overall conversion readiness
score...."""

USER_MESSAGE_TEMPLATE = (
    "Analyze this e-commerce store's UX and conversion readiness: {url}\n\n"
    "Browse product pages, attempt the checkout flow, and check mobile layout."
)

REPORT_HEADER_TAG = "UX & Conversion Assessment"

SaaS Security Compliance

Assess a SaaS product's public posture against SOC 2 and ISO 27001 requirements.

SYSTEM_PROMPT = """You are a cybersecurity analyst specializing in
SaaS compliance. Evaluate the target company's public security posture:
trust/security page, SOC 2 Type II report availability, data encryption
mentions, SSO/MFA support, GDPR/DPA documentation, incident response
policy, and subprocessor transparency. Score readiness 1-10...."""

USER_MESSAGE_TEMPLATE = (
    "Evaluate this SaaS product for SOC 2 / ISO 27001 readiness: {url}\n\n"
    "Check their security page, docs, trust center, and terms of service."
)

REPORT_HEADER_TAG = "Security Compliance Assessment"

License

MIT