Metadata-Version: 2.4
Name: idcanopy-aletheia-mcp
Version: 0.1.0
Summary: MCP server for Aletheia — IDCanopy's document-fraud detection for AI agents
Author: IDCanopy / CINDR.LA
License: Proprietary
Project-URL: Homepage, https://forensic.s1.cindrla.world/portal/
Keywords: mcp,forensics,fraud-detection,kyc,document-analysis
Classifier: Programming Language :: Python :: 3.12
Classifier: Operating System :: OS Independent
Classifier: Intended Audience :: Financial and Insurance Industry
Requires-Python: >=3.12
Description-Content-Type: text/markdown
Requires-Dist: mcp>=1.0

# idcanopy-aletheia-mcp

MCP server for the **IDCanopy forensic-scoring-api** — a document-fraud-detection
service for bank onboarding. Exposes the analysis API as MCP tools that an AI
agent (Claude Desktop, Cline, codexx, …) can call to score applicant documents
for fraud and pull back auditable findings and PDF reports.

This package is **standalone**: it depends only on the [`mcp`](https://pypi.org/project/mcp/)
SDK and the Python standard library — not on the forensic engine itself. It talks
to a running forensic-scoring-api instance over HTTP.

## Install

```bash
pip install idcanopy-aletheia-mcp
```

## Run

```bash
export FORENSIC_API_TOKEN=<bearer>                    # required; env only, never logged
export FORENSIC_API_BASE_URL=https://forensic.s1.cindrla.world
idcanopy-aletheia-mcp
```

Or wire it into an MCP client (`~/.claude/mcp.json`, Cline, codexx — see the full
config snippets in the project portal at <https://forensic.s1.cindrla.world/portal/>):

```json
{
  "servers": {
    "forensic": {
      "command": "idcanopy-aletheia-mcp",
      "env": {
        "FORENSIC_API_TOKEN": "<bearer>",
        "FORENSIC_API_BASE_URL": "https://forensic.s1.cindrla.world"
      }
    }
  }
}
```

## Tools

| Tool | Purpose |
|---|---|
| `analyze_documents` | Stateless one-shot analysis of document files → verdict + findings |
| `create_applicant` | Create a persistent applicant record |
| `upload_documents` | Attach documents to an applicant |
| `get_applicant_result` | Analyze an applicant's stored documents and return the result |
| `get_findings` | Re-read a stored result without re-analyzing |
| `analyze_batch` | Submit an async batch job over many applicants |
| `get_batch_status` | Poll a batch job |
| `get_report` | Generate the branded PDF report to a local file |

`analyze_documents` and `get_applicant_result` emit MCP progress notifications on
long calls when the client supplies a `progressToken`.

## Security

The bearer token is read from `FORENSIC_API_TOKEN` only. It is never logged,
echoed, or included in any tool's return value or error text.

## License

Proprietary — © IDCanopy / CINDR.LA.
