# forency [Health: Active]

**Category:** 🛒 E-Commerce  
**Repository:** https://github.com/cheevahagadog/forency-mcp  
**GitHub Stars:** 1  
**npm Downloads (last month):** 116  
**Views:** 1  
**Installs:** 0  
**Upvotes:** 0  
**Directory Page:** https://allmcps.com/mcp/forency

## Description
Detect the tech stack behind any website: CMS, ecommerce, analytics, CRM, payments, hosting.

## Tools
Capabilities this server exposes over MCP:

- **lookup_tech_stack** — Detect the technologies a website runs: CMS, ecommerce platform, JavaScript frameworks, analytics, CRM, marketing automation, payment processors, chat widgets, CDN, and hosting. Use it to research a competitor, qualify a sales prospect, or check what a site is built with. Reads the HTML and headers the server sends, so anything injected only by client-side JavaScript may not appear — a sparse result means little was detectable, not that the site runs nothing. Counts as one scan against the caller's plan.
- **batch_lookup** — Detect the technology stacks of up to 25 websites in one call, returning the full detail for each. Use it when working through a list of prospects, competitors, or portfolio sites. If you want a side-by-side view of who uses what, use compare_stacks instead. Counts one scan per URL against the caller's plan.
- **compare_stacks** — Scan 2-25 websites and return a side-by-side table of which technologies each one uses. Answers questions like 'which of these prospects use HubSpot vs Salesforce?' or 'how does our stack compare to our competitors?'. For the full per-site detail rather than a table, use batch_lookup. Counts one scan per URL against the caller's plan.

## Claude Desktop Quick Installation
Install path detected from listing signals. Uses `npx` (confidence: high):

```json
"mcpServers": {
  "forency": {
    "command": "npx",
    "args": ["-y","forency-mcp"],
    "env": {
      "FORENCY_API_KEY": ""
    }
  }
}
```

**Requires environment variables:** `FORENCY_API_KEY` — the values above are empty placeholders; fill in real credentials before running (see the repository for what each one is for).

## Documentation

## What the forency MCP server does

The forency MCP server lets an AI assistant investigate the technologies associated with websites. Results can cover CMS platforms, ecommerce systems, JavaScript frameworks, analytics, CRM, marketing automation, payment services, chat widgets, CDNs, and hosting. The service is useful for prospect research, competitive analysis, and comparing the technical choices visible across several sites.

It provides three operations:

- `lookup_tech_stack(url)` examines one website and returns a detailed technology breakdown.
- `batch_lookup(urls)` processes as many as 25 websites in one request.
- `compare_stacks(urls, categories?)` creates a side-by-side comparison and can restrict the result to categories such as CRM or Analytics.

Each URL counts as one scan against the account plan.

## How it works

The local package is a thin stdio MCP client. It receives tool calls from an MCP-enabled application, sends the requested URLs and API key to the Forency REST API, and formats the response for the client. By default, requests go to `https://forency.io`; the API base can be changed with an environment variable.

Detection takes place remotely rather than in the package. Forency examines the HTML and HTTP headers returned by a site. It does not operate a browser, so technologies added only after client-side JavaScript runs may not be detected. A limited result therefore indicates limited detectable evidence, not necessarily an absence of technology.

The package documentation states that it sends the API key and requested URLs to Forency, without telemetry, analytics, or third-party calls from the client itself.

## Setup and configuration

Local installation uses `npx`, so add the server to the MCP configuration used by the client. Set `FORENCY_API_KEY` to a key from forency.io. Claude Code can add the same server with its MCP command and an environment variable assignment.

Remote-capable clients can connect directly to `https://forency.io/api/mcp` and provide the API key as a Bearer token. This avoids installing the local package and uses the hosted tools.

The optional `FORENCY_API_URL` variable changes the API destination from the default Forency address. This is useful when you need to observe or redirect the requests made by the client. A free account includes 100 scans per month, according to the provided pricing information.

## Tools and capabilities

The forency MCP server supports individual inspection, batch processing, and comparisons. Batch requests are capped at 25 websites per call. Comparison requests accept an optional list of technology categories, allowing an assistant to focus on selected signals instead of presenting every detected category.

The returned information depends on what can be identified from a site's server response. The client package contains no fingerprint database or detection implementation; the detection service relies on technology fingerprints maintained by the webappanalyzer project.

## Limitations and notes

A Forency API key is required for local use and for authenticated hosted access. Usage is measured per scanned URL, including URLs included in batch and comparison calls.

Because analysis is based on returned HTML and headers rather than browser execution, client-side-only technologies can be missed. The local client communicates with the configured API endpoint and does not perform detection independently. The package is released under the MIT license.

_Full upstream README: https://allmcps.com/mcp/forency/readme_