# attention.run

> attention.run recognises buttons, fields and dialogs from pixels in milliseconds, with no AI model call per step, and automates Chrome and any desktop over VNC. Invite-only preview at https://app.attention.run.

attention.run recognises UI elements by deterministic template matching (a native engine, cached by the pixels of the screen) and drives Chrome over CDP and any desktop over VNC. People record sessions, cut a library of elements from them, and automate with JSON flows, a TypeScript SDK, or AI agents over MCP. The docs below cover each, with examples that are run in CI against a local hub.

## Product

- [attention.run — Automate anything you can see](https://attention.run/index.md): attention.run recognises buttons, fields and dialogs from pixels in milliseconds, with no AI model call per step, and automates Chrome and any desktop over VNC.
- [Request access](https://attention.run/request-access.md): attention.run is in an invite-only preview. Tell us what you’d automate; we invite people in small batches.

## Docs

- [attention.run docs](https://attention.run/docs.md): How to use attention.run: from a local hub to a flow that passes, the flow JSON reference, the attn CLI, the TypeScript SDK, MCP, SSH targets and webhooks.
- [Getting started](https://attention.run/docs/getting-started.md): From nothing to a passing flow: start a local hub with the fixture app, record a session, cut four elements, write a flow and run it from the CLI.
- [Concepts](https://attention.run/docs/concepts.md): The parts of attention.run and how they fit: targets, sessions and recordings, elements and libraries, flows, runs and triggers, and the hub and runners.
- [Flow JSON reference](https://attention.run/docs/flow-json.md): Every field and step of an attention.run flow: selectors, targets, conditions, params, outputs and handlers, generated from the schema the hub validates with.
- [The attn CLI](https://attention.run/docs/cli.md): The attn command: push flows, start runs and wait for them in CI with exit codes, read a run's steps and outputs, share a screen, and run MCP for local agents.
- [The TypeScript SDK](https://attention.run/docs/sdk.md): @attn/sdk, the hub API from TypeScript: push flows, start runs and wait for them, drive a live session call by call, and share a screen from a script.
- [MCP](https://attention.run/docs/mcp.md): attention.run's MCP tools for AI agents: open a session, read the screen as text, act and wait, run flows, and save what the agent did as a flow.
- [SSH targets](https://attention.run/docs/ssh-targets.md): A desktop or a Chrome on your own machine as an attention.run target: add an SSH key, run one ssh -R command or attn share, and stop it with Ctrl+C.
- [Webhooks](https://attention.run/docs/webhooks.md): Start attention.run runs from other systems with a webhook trigger, and get a signed request when a run ends: payload, headers, retries and signatures.

## Company

- [Privacy policy](https://attention.run/privacy.md): What attention.run collects on this site, through the request form and in the app, why, where it is kept and for how long.
- [Terms of the preview](https://attention.run/terms.md): The terms for using attention.run during its invite-only preview.

## Optional

- [Everything above in one file](https://attention.run/llms-full.txt)
- [The flow JSON Schema](https://attention.run/schema/flow-v1.json): what the hub validates flows with