The Angelcam MCP Server connects Claude, ChatGPT and other AI assistants to your Angelcam account. Ask in plain language whether a camera works right now, what it recorded at 3am, or what needs attention across a site, and your assistant answers from your own Angelcam data.
It works through the Model Context Protocol (MCP), the open standard AI assistants use to call tools in other services. You add one URL to your assistant, sign in with your Angelcam account, and that's it. The MCP Server is in beta and free on top of your Angelcam subscription.
What you can ask
Ask the way you would ask a colleague. The assistant picks the right tools.
"Is the Front Door camera actually working right now?" The assistant pulls a fresh frame from the camera, not just the last status Angelcam recorded.
"Brief me on what needs attention at the site this morning." Offline cameras, stale snapshots, recording problems and incidents across the space.
"What did the loading bay camera record at 3am?" Frames from the footage around that moment, plus a link to play it back.
"Which cameras have gaps in their recordings from the last 24 hours?" Recorded segments and real gaps for each camera.
"Has any camera been moved or had its view blocked?" For cameras with defect detection set up, the result of each check and why it was reached.
"What Real Time Security incidents happened overnight?" What raised each incident, whether it is still open, and how it was verified.
"Start recording on the Warehouse camera." Cloud Recording starts once you confirm the camera the assistant names.
If your assistant also has other tools connected, such as Slack or Asana, it can turn the answer into a message, a task or a daily briefing. Those actions happen in those tools, under their own permissions, not in Angelcam.
What it can and cannot do
It reads: your spaces, locations and cameras, live and stored snapshots, live stream links, Cloud Recording (timelines, gaps, past frames, playback and download links), Real Time Security incidents, defect detection checks, and active services.
It can change exactly two things:
Start or stop Cloud Recording on a camera. The assistant is instructed to confirm with you first and name the camera.
Send feedback about a bug or a missing feature to the Angelcam team, only after you have agreed to what the report contains.
It never arms or disarms locations, changes camera settings, or deletes cameras, recordings or anything else.
Want a connection that can never change anything? Use the read-only URL below. It doesn't offer the recording or feedback tools at all, and Angelcam itself refuses a recording change made through it. This is the safer choice for a connector an admin adds for a whole organization, or for a cautious first try.
Connect your assistant
Pick one of two URLs:
Full connector: https://mcp.angelcam.com (reads, plus recording control and feedback)
Read-only connector: https://mcp.angelcam.com/readonly (reads only)
Claude (Free, Pro, Max): Customize → Connectors, click + then Add custom connector, paste the URL, then Add. The Free plan allows one custom connector.
Claude (Team, Enterprise): an Owner adds it once under Organization settings → Connectors → Add → Custom → Web. Members can't add it themselves; once it's there, each member finds it under Customize → Connectors (marked Custom) and clicks Connect.
ChatGPT (paid plans, web app): turn on Developer mode under Settings → Apps → Advanced settings (OpenAI moves this setting from time to time), then create a new app, paste the URL as the MCP server URL and choose OAuth. On Business and Enterprise workspaces an admin may need to enable developer mode first.
Claude Code: run claude mcp add --transport http angelcam https://mcp.angelcam.com
Cursor, VS Code, Windsurf and other clients: one-click install buttons and config snippets are on mcp.angelcam.com. Any client that supports remote MCP servers over Streamable HTTP with OAuth works.
When you connect, you first see a consent page. It shows which app is connecting and the permissions it will get, each with a checkbox, so you can untick the optional ones. Then you sign in with your Angelcam account (email and password, Google or Apple sign-in all work). Your password is never shared with the AI assistant.
Security and privacy
Your account, your permissions. Every request uses your own Angelcam sign-in, so the assistant sees exactly what you see in MyAngelcam and nothing more.
Secure sign-in. The connection uses OAuth 2.1 with PKCE. To stop your assistant from using your account, disconnect the connector in your assistant.
Minimal logging. Audit logs record which tool ran and how it ended. They never record your prompts, email address, tokens or stream links.
Snapshots go to your AI provider. Snapshots you ask for are sent to your assistant as pictures, so the company that runs it (for example Anthropic or OpenAI) receives them under your own terms with that company.
Good to know
Spaces. If your account belongs to several spaces, each with its own cameras, the assistant asks which one you mean.
Shared cameras. Cameras another account shared with you live in that account's space, not yours. Tell the assistant a camera is shared with you, or ask "Which cameras are shared with me?"
Times are in UTC. Ask in your own time zone; the assistant converts it.
Recordings are read in windows. One lookup covers up to 24 hours of a recording timeline, and a download link up to 3 hours. For longer stretches the assistant splits the request.
Recordings need Cloud Recording. Footage questions, and starting or stopping recording, only work for cameras with Cloud Recording.
About 30 requests a minute. Most questions take one to three. A long review may pause for a few seconds.
Snapshots in the chat. In Claude and ChatGPT a requested snapshot appears under the answer, with the camera name and time. Other assistants describe it instead.
Troubleshooting
My assistant says the connector is not authorized. The sign-in usually expired or wasn't finished. Disconnect the connector in your assistant, connect it again and complete the Angelcam sign-in.
Starting or stopping a recording says I don't have permission. Check three things. Connections made before 2 October 2026 don't include recording permission: disconnect and connect once more. You may be using the read-only URL. Or you unticked "Starting and stopping recording" or "Cloud recordings" on the consent page; recording control needs both.
I see no cameras, or the wrong ones. Ask the assistant to list your spaces and tell it which one to use. A camera missing in MyAngelcam is missing here too.
Still not working? Check status.angelcam.com for an outage, then ask your assistant to send feedback to Angelcam, or email us.
FAQ
Does it cost anything? No. It's included with your Angelcam account and uses the services you already have.
How is this different from the Angelcam API? The Angelcam API is for code you write yourself. The MCP Server exposes the same platform as tools an AI assistant can use directly, under your own sign-in.
Does it replace the MyAngelcam dashboard? No. MyAngelcam remains the place to manage your cameras and settings. The MCP Server makes everyday questions faster to answer.
Can I rely on the answers? AI assistants can misread data. For security, compliance or incident response decisions, confirm important findings in MyAngelcam before acting.
The full tool list and setup options are kept up to date on mcp.angelcam.com.
😇 We are here to help
Found a bug or missing a feature? Ask your assistant to send feedback, or just say hello@angelcam.com.