Documentation

Everything you need to connect an agent to live visual verification data.

Quick start

Vancam exposes the same backend through three interfaces: MCP (recommended for Cursor and Claude), CLI (scripting, cron jobs, quick lookups), and REST (for custom agents and backends). All three require an API key — get one free from your account dashboard.

Prerequisites

  • Python 3.10+ (for the local MCP server or CLI)
  • Clone mcp-vancam and install dependencies

MCP setup

git clone https://github.com/shughestr/mcp-vancam.git
cd mcp-vancam
pip install -r requirements.txt

Add the server to your MCP client config. See the MCP page for tool reference and client examples.

CLI setup

cd vancam-cli
pip install -e .
export VANCAM_API_KEY=your_api_key_here

See the CLI page for the full command reference.

REST setup

No SDK required. Call the endpoints directly with an x-api-key header. Download the spec from openapi.yaml or read the API reference.

Core concepts

Assets

Each feed source has an asset_id. Responses include coordinates, metadata, and a live image_url.

Search modes

  • Radius — feeds near a point (lat, lon, radius in km)
  • Nearest — N closest feeds to a point
  • Route — feeds along a path with buffer
  • Bounds — feeds inside a bounding box

Agent discovery

Machine-readable index: llms.txt

Related repos