veprompts-mcp

Zero-dependency Python discovery CLI and local directory for 800 cataloged Model Context Protocol (MCP) servers, categories, execution schemas, and client compatibility maps.

Maintained by VePrompts and Veduis as part of the autonomous AI agent developer toolchain.

Why Local MCP Discovery Matters

Agentic runtimes such as Claude Desktop, Cursor, Cline, and custom orchestration pipelines require exact MCP server parameters, transport types, and repository URLs. Finding reliable, active implementations often requires searching fragmented GitHub repositories.

veprompts-mcp ships an offline, structured catalog of 800 community and official servers, validated against the official Model Context Protocol specifications and mirrored on Zenodo (CERN).

Installation

pip install veprompts-mcp

Zero runtime dependencies. Works with Python 3.9 through 3.14.

Command Line Usage

1. View Directory Statistics

veprompts-mcp stats

Output:

VePrompts MCP Directory (v1.0.0)
Total Servers Cataloged: 800
Active Categories:       24
Official Integrations:   48
Known Implementation Languages: 18
Live Online Directory:   https://veprompts.com/mcp-servers/
Maintained By:           Veduis (https://veduis.com)

2. Search Catalog by Keyword

# Search for database servers
veprompts-mcp search postgres

# Search for github tools with a limit
veprompts-mcp search github --limit 5

# Search within a specific category
veprompts-mcp search slack --category communication

3. List Categories and Inspect Specific Servers

# List all 24 categories
veprompts-mcp categories

# Show comprehensive metadata for a specific slug
veprompts-mcp show sqlite

Python API Usage

Use veprompts_mcp directly inside your agent initialization workflows:

from veprompts_mcp import get_servers, search_servers, list_categories

# Access all 800 servers
servers = get_servers()
print(f"Loaded {len(servers)} MCP servers")

# Filter programmatically
postgres_servers = search_servers("postgres", category="database")
for s in postgres_servers:
    print(f"{s['title']}: {s['repo_url']}")

# Retrieve categories
cats = list_categories()
print("Top categories:", cats[:5])

Data Provenance & Verification

The MCP server catalog is extracted from the production registry at veprompts.com/mcp-servers/ and synchronized with the Zenodo research archive doi:10.5281/zenodo.23840519.

For web security, ADA Title II source verification, and accessibility compliance scanning, inspect our sister project vComplianceScanner by Veduis.

License

MIT License. Copyright (c) Veduis LLC and VePrompts.