mcpField
S
Spot AI MCP (Unofficial)

Spot AI MCP (Unofficial)

Local
github.com/phil-bryantSecurity0Updated 8w ago
Repository

About Spot AI MCP (Unofficial)

Spot AI MCP is an unofficial, read-only MCP server for querying Spot AI camera and video-intelligence data. It exposes the API through stdio tools and cannot modify an organization.

Key features

  1. 01Browse locations and cameras
  2. 02View camera status and details
  3. 03Query people, vehicle, and forklift events
  4. 04Retrieve license plate reports
  5. 05Generate live-stream viewing URLs
  6. 06Read documented API paths

Use cases

  1. 01Inventory accessible locations and cameras
  2. 02Inspect zones configured on a camera
  3. 03Review video-intelligence events by date range
  4. 04Get a camera's live-stream viewing URL

MCP support

The server exposes read-only queries to the Spot AI REST API as MCP tools. It communicates with MCP clients over stdio and marks its tools as read-only.

Configuration

Local server config is the same for every client. Copy it into your client’s MCP config file.

mcp.json · Universal config
{
  "mcpServers": {
    "phil-bryant-spot-ai": {
      "command": "uvx",
      "args": [
        "spot-ai-mcp"
      ],
      "env": {
        "SPOT_AI_API_KEY": "<SPOT_AI_API_KEY>"
      }
    }
  }
}
Run commanduvx spot-ai-mcp
Environment variables to set
SPOT_AI_API_KEYSecretSpot AI API key (create one in the Spot AI dashboard API settings and add an authorization role to it)

Local servers need a runtime on your machine (e.g. Node.js 18+). Config file locations differ by client; see your client’s docs.

Packages
pypispot-ai-mcp0.3.0

Tools

10

Saved tool list · updated 2w ago

ParameterTypeRequiredDescription
limitnumberNoMax results per page
cursorstringNoPagination cursor from a previous response's 'next' field

Details

From the project README

spot-ai-mcp

Unofficial community MCP server for the Spot AI camera / video-intelligence REST API. Not affiliated with or endorsed by Spot AI.

Single-file, stdio transport, no dependencies beyond Python 3.9+. Read-only by design: it can browse cameras and intelligence data but can never modify anything in your Spot AI org. Every tool wraps a GET endpoint except get_live_stream_urls, which wraps POST /v1/cameras/live — a read-like POST that only generates a viewing URL. All tools declare the readOnlyHint: true MCP annotation.

Install

uvx spot-ai-mcp

or pip install spot-ai-mcp, or run straight from a checkout with python3 -m spot_ai_mcp (no dependencies to install).

Register with Claude Code:

claude mcp add spot-ai -s user -e SPOT_AI_API_KEY=YOUR_KEY -- uvx spot-ai-mcp

API key

Create a key in the Spot AI dashboard's API settings, then add an authorization (a role, e.g. Owner, optionally scoped) on the key's settings page. A key without a role returns empty lists from every resource endpoint while get_camera_count still works — that's the tell.

The server resolves the key lazily on the first API call:

  1. SPOT_AI_API_KEY environment variable — the normal path.
  2. Optionally, a secret-helper command, so the key never sits in an env var or config: the server runs $SPOT_AI_OP_BIN -f $SPOT_AI_OP_ITEM $SPOT_AI_OP_FIELD (defaults 1psa -f spot.ai api_key, per 1psa, a vault-scoped 1Password service-account CLI). Point these at any command with the same flag convention.

The key is never written to disk or config by this server.

Tools

ToolWhat it does
list_locationsLocations the key can see (paginated)
list_camerasCameras with status, location, IP, MAC (paginated)
get_cameraOne camera by id
get_camera_countNumber of enabled cameras in the org
list_appliancesIntelligent Video Recorders (paginated)
get_zonesZones defined on a camera
get_intelligenceCounting / idle / presence events for people, vehicles, or forklifts over a date range
get_lpr_reportLicense-plate-recognition report for an LPR camera
get_live_stream_urlsLive-stream viewing URL for up to 4 cameras
spot_api_getEscape hatch: GET any documented /v1/ or /v2/ path

Notes

  • Dual-era MCP server: speaks both the modern per-request protocol (server/discover, spec 2026-07-28) and the legacy initialize handshake (2024-11-05 through 2025-06-18), so old and new clients both work.
  • Base URL is https://dev-api.spot.ai, auth is Authorization: Bearer <key>.
  • Cloudflare in front of the API rejects Python's default user agent with error 1010; the server sends User-Agent: spot-ai-mcp/<version>.
  • Endpoint index: https://developers.spot.ai/llms.txt (append .md to any docs URL for markdown, including the OpenAPI definition per endpoint).

License

MIT

Comments

Markdown supported

New to MCP?

From what the protocol is to installing your first server, these answers cover most first questions.

An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.