ATM · Advanced Token Management for Home Assistant

Why ATM instead of a long-lived access token

A long-lived access token (LLAT) plus the native MCP server gives every client the same all-or-nothing view of your home. ATM keeps the same tools and adds the control layer that the native system has no place for.

Capability	LLAT + native MCP	ATM token
MCP tool compatibility	20 native tools	Same 20 tools, identical names and responses, plus 78 more
Entity filtering	Binary: expose or hide, same for all clients	Four permission states, per token
Per-client control	No, all clients share one exposed set	Yes, independent permissions per token
Read-only access	No	Yes, READ allows reads and blocks writes
Audit trail	None	Every request logged with outcome and entity
Rate limiting	None	Per token, configurable
Expiry	None	Optional, auto-archived on expiry
Revocation	Revoke the LLAT on the HA profile page	Instant; the token's next request is rejected
Sensitive attribute scrubbing	None	Always applied
Client reconfiguration	`/api/mcp`	`/api/atm/mcp` (URL change only)

If you are connecting Claude Code, Cursor, ChatGPT, Antigravity, or any other AI tool to Home Assistant, ATM gives you control the native system cannot.

How ATM compares:

We benchmarked ATM against Home Assistant's built-in MCP server and the popular community server, one agent model, one synthetic home, the same tasks. ATM came out cheapest per completed task, the only server that left every off-limits device untouched, and it's the most capable.

Results

Read the road test

Cost, capability, and safety across the three servers, with charts and the full method.

View the report Reproduce

Run it yourself

The harness, tasks, and raw run data are open. Point it at your own Home Assistant and check the numbers.

View the repository

Three layers between an agent and your home

Every request runs the same gauntlet, in order. ATM resolves and flattens the target to explicit entities, checks the token's permissions and capabilities, and only then lets MESA, the semantic safety net, have the final say, a backstop that holds even over a token that was granted too much.

The permission tree

Domains, devices, and entities, each set to read-write, read-only, deny, or inherit. A two-pass resolver decides what every request can touch.

Permissions

Capability flags

Twenty-two opt-in switches gate the high-impact operations: restart, physical control, writing automations, reading logs, editing YAML. Off by default.

Capabilities

Semantic safety profiles

MESA stores a semantic profile for each entity: what it is, how sensitive it is, and how it should be treated by your AI agent, not just allow or deny. Describe your home's nature once, and every agent inherits the safe treatment: a lock or camera stays read-only or confirmation-only because of what it is, no matter which token is asking.

MESA

From install to a scoped agent in four steps

01 · Install

Install & set up

Add ATM through HACS or by hand, restart, and add the integration. One ATM instance per Home Assistant.

Installation 02 · Connect

Quick start

The guided setup creates your first token, gives it one device, connects your agent, and tests it.

Quick start 03 · Scope

Scope its access

Paint the permission tree, select whole areas at once, and verify the outcome with the effective-permission emulator.

Permissions 04 · Operate

Know the tools

Eighty-six tools, from get_state to create_automation, each scoped to the token and its capabilities.

Tools reference

Prefer not to install a client? Chat inside Home Assistant

Agent Chat runs the agent inside ATM itself, on a token you choose, using your own Claude, DeepSeek, Ollama or other provider account. Same permission tree, same capabilities, same approvals and MESA, just a chat window floating above Home Assistant instead of an external tool.

Internal