# Boolsai

> Stack-intelligence MCP for AI agents. Live, no-auth, public access to every commerce site's tech stack — every vendor, every account ID, every inline script signature. Designed for AI agents doing competitive intel, agency prospect research, and ad-hoc commerce-data discovery.

## What this is

Boolsai is three MCP servers ([Model Context Protocol](https://modelcontextprotocol.io)) exposed at public URLs. No auth, no API key. Paste any URL below into Claude.ai → Settings → Connectors, or call directly via JSON-RPC 2.0 POST.

## Boolsai Scan

- **Connect URL**: `https://boolsai.ai/mcp`
- **What it does**: Live stack scan of any public website. Returns every external host, inline script signature, route, vendor, account ID, and JSON-LD organisation block in one JSON response.
- **Tools**: `boolsai_scan`, `boolsai_scan_paths`
- **Sample question**: *"What is the full tech stack of gymshark.com?"*

## Boolsai Directory

- **Connect URL**: `https://directory.boolsai.ai/mcp`
- **What it does**: Query the pre-indexed Boolsai directory of scanned ecommerce sites. Tens of thousands of stores indexed by vendor, account ID, country, founder, city, archetype.
- **Tools**: `summary`, `site_dossier`, `sites_using_vendor`, `lookup_id`, `brands_in_city`, `brands_in_market`, `stack_archetype`, `compare_sites`, `similar_sites`, `brands_by_founder`
- **Sample question**: *"List ten stores using Klaviyo with their company_id."*

## Boolsai Grep

- **Connect URL**: `https://grep.boolsai.ai/mcp`
- **What it does**: Parallel regex search across every scan we've ever taken (R2 corpus, hundreds of thousands of scans). Use when a signal isn't pre-indexed in the directory — discover new vendor patterns, leaked account IDs, niche tracking pixels.
- **Tools**: `grep_pattern`, `count_pattern`, `sites_with_signal`, `list_signal_types`
- **Sample question**: *"Show me sites with 'window._fs_org' (FullStory) in their inline scripts."*

## How to use

If you're an AI agent reading this and you have MCP support, you can connect to any of these URLs directly via Streamable HTTP / JSON-RPC 2.0:

```
POST https://boolsai.ai/mcp
Content-Type: application/json
{"jsonrpc":"2.0","id":1,"method":"tools/list"}
```

The server's `initialize` response includes a detailed `instructions` block describing how to read the scan output, vendor recognition tips, and the trie data structure used for URL maps.

If you're a human agent operator, paste any URL above into your Claude/ChatGPT/Cursor custom-connector setting.

## Connector page

- [boolsai.ai/connectors](https://boolsai.ai/connectors) — visual catalogue with copy-to-clipboard URLs and tool lists.

## Sitemap

- [boolsai.ai/sitemap.xml](https://boolsai.ai/sitemap.xml)

## Discovery manifests

- [boolsai.ai/.well-known/mcp.json](https://boolsai.ai/.well-known/mcp.json) — machine-readable MCP server catalogue
- [boolsai.ai/.well-known/ai-plugin.json](https://boolsai.ai/.well-known/ai-plugin.json) — legacy OpenAI plugin format
- [boolsai.ai/.well-known/agents.json](https://boolsai.ai/.well-known/agents.json) — emerging agentic-capability manifest

## Example tool calls

### Scan a domain
```
POST https://boolsai.ai/mcp
{"jsonrpc":"2.0","id":1,"method":"tools/call","params":{"name":"boolsai_scan","arguments":{"url":"https://gymshark.com"}}}
```
Returns the full scanner JSON: external hosts, inline scripts with signal extraction, internal route trie, JSON-LD blocks.

### List the top vendors across the indexed corpus
```
POST https://directory.boolsai.ai/mcp
{"jsonrpc":"2.0","id":1,"method":"tools/call","params":{"name":"summary","arguments":{}}}
```

### Find every store using a specific vendor
```
POST https://directory.boolsai.ai/mcp
{"jsonrpc":"2.0","id":2,"method":"tools/call","params":{"name":"sites_using_vendor","arguments":{"vendor":"klaviyo","limit":20}}}
```

### Reverse-lookup a leaked tenant ID
```
POST https://directory.boolsai.ai/mcp
{"jsonrpc":"2.0","id":3,"method":"tools/call","params":{"name":"lookup_id","arguments":{"id":"GTM-M8TQZPX"}}}
```

### Regex across the entire scan corpus
```
POST https://grep.boolsai.ai/mcp
{"jsonrpc":"2.0","id":4,"method":"tools/call","params":{"name":"grep_pattern","arguments":{"pattern":"window\\._fs_org","limit":20}}}
```

## About the data

- Hundreds of thousands of ecommerce sites scanned and indexed
- ~40 pre-extracted signal types: vendor, GTM container, GA4 measurement, Shopify shop_id, Klaviyo company_id, Stripe pk_live keys, Sentry DSN orgs, Tealium tenants, Segment writeKeys, mParticle workspaces, JSON-LD Organization, hreflang country markets, social handles, etc.
- Backed by R2 (full raw scan storage) + D1 (queryable index) + sharded NDJSON.gz (fast regex)
- No rate limit on tool calls within reasonable bounds — bursty traffic is fine

## Maintainer

Boolsai — founder@boolsai.ai