# Modal Computer Use

> Deploy and control a Linux desktop in a Modal Sandbox.

- [Build computer-use agents on Modal](https://modal-computer-use.mintlify.app/index.md): Run a Linux desktop in a Modal Sandbox and control it with Python.
- [Install Modal Computer Use](https://modal-computer-use.mintlify.app/start/installation.md): Install the Python package and set up Modal credentials.
- [Run your first placed trajectory](https://modal-computer-use.mintlify.app/start/quickstart.md): Create one desktop and lend it to one placed Modal Function.
- [Choose an example](https://modal-computer-use.mintlify.app/start/examples.md): Choose a maintained example for the placed or low-level SDK path.
- [Create, attach, and borrow](https://modal-computer-use.mintlify.app/build/create-and-attach.md): Choose the lifecycle and cleanup contract for one Modal desktop.
- [Configure browser automation](https://modal-computer-use.mintlify.app/build/browser-automation.md): Choose a browser, Image source, prewarm policy, and resources.
- [Run ordered action batches](https://modal-computer-use.mintlify.app/build/action-batches.md): Validate and run one ordered model action array without automatic replay.
- [Capture screenshots and recordings](https://modal-computer-use.mintlify.app/build/screenshots-recordings.md): Capture semantic screenshots, choose a source, and record the desktop.
- [Store and persist artifacts](https://modal-computer-use.mintlify.app/build/artifacts-storage.md): Read, write, download, and persist files under the artifact root.
- [Hand a desktop to a Modal Function](https://modal-computer-use.mintlify.app/build/session-handoff.md): Pass a versioned session handle to one explicitly placed Function.
- [Build an application run gateway](https://modal-computer-use.mintlify.app/build/run-gateway.md): Expose a bounded spawn-and-poll control plane from your application.
- [Run an OpenAI computer loop](https://modal-computer-use.mintlify.app/integrate/openai.md): Connect the OpenAI Responses API computer tool to a placed Modal desktop.
- [Run an Anthropic computer loop](https://modal-computer-use.mintlify.app/integrate/anthropic.md): Connect the Anthropic Messages API computer tool to a placed Modal desktop.
- [Deploy a placed Modal trajectory](https://modal-computer-use.mintlify.app/operate/deploy.md): Deploy one Function and one owned desktop with explicit placement and cost.
- [Configure placement, cost, and limits](https://modal-computer-use.mintlify.app/operate/configure.md): Set explicit placement, resources, Image policy, input limits, and warm capacity.
- [Keep warm desktop capacity](https://modal-computer-use.mintlify.app/operate/warm-capacity.md): Fill and claim a bounded pool of ready Modal desktop Sandboxes.
- [Measure and reduce trajectory latency](https://modal-computer-use.mintlify.app/operate/performance.md): Separate startup from warm work and measure the placed action-to-frame path.
- [Secure a desktop deployment](https://modal-computer-use.mintlify.app/operate/security.md): Protect daemon access, noVNC, provider loops, logs, and retained artifacts.
- [Troubleshoot a desktop](https://modal-computer-use.mintlify.app/operate/troubleshooting.md): Diagnose readiness, X11 input, screenshots, adapters, storage, and warm capacity.
- [Benchmark overview](https://modal-computer-use.mintlify.app/benchmarks/overview.md): Find current results, evidence definitions, and benchmark commands.
- [Current benchmark results](https://modal-computer-use.mintlify.app/benchmarks/current-results.md): Review current and dated benchmark measurements with their scope.
- [Weighted input-capacity result](https://modal-computer-use.mintlify.app/benchmarks/input-capacity.md): Review the evidence behind the 100-token refill and 400-token burst defaults.
- [Latency evidence](https://modal-computer-use.mintlify.app/benchmarks/latency-evidence.md): Map each performance claim to its status, measurement boundary, and proof.
- [Run benchmarks](https://modal-computer-use.mintlify.app/benchmarks/run.md): Run credential-free checks and authorized Modal benchmarks.
- [Reproduce a benchmark](https://modal-computer-use.mintlify.app/benchmarks/reproducibility.md): Verify tracked evidence and publish a new dated result.
- [API reference overview](https://modal-computer-use.mintlify.app/reference/overview.md): Choose the Python SDK or daemon API surface for your application.
- [Migrate from v1 to v2](https://modal-computer-use.mintlify.app/reference/migration-v2.md): Move a v1.1.0 application to the v2.0.1 ownership, placement, and Step contracts.
- [Configuration reference](https://modal-computer-use.mintlify.app/reference/configuration.md): Configure Modal orchestration with ComputerConfig and a direct daemon with environment variables.
- [Namespace reference](https://modal-computer-use.mintlify.app/reference/namespaces.md): Find Python methods for Step, input, screenshots, artifacts, and daemon capabilities.
- [Errors and models](https://modal-computer-use.mintlify.app/reference/errors-and-models.md): Review stable result models and recovery errors.
- [OpenAPI reference](https://modal-computer-use.mintlify.app/reference/openapi.md): Use and verify the checked-in OpenAPI 3.1 schema.
- [Control a Linux desktop on Modal](https://modal-computer-use.mintlify.app/v1/index.md): Create and control a Linux desktop in a Modal Sandbox with Modal Computer Use.
- [Install Modal Computer Use](https://modal-computer-use.mintlify.app/v1/start/installation.md): Install modal-computer-use and set up Modal credentials.
- [Create your first desktop](https://modal-computer-use.mintlify.app/v1/start/quickstart.md): Create a browser desktop, open a page, and save a screenshot.
- [Choose an example](https://modal-computer-use.mintlify.app/v1/start/examples.md): Choose a maintained Modal Computer Use example for your lifecycle, provider, or storage use case.
- [Create and attach to desktops](https://modal-computer-use.mintlify.app/v1/build/create-and-attach.md): Choose a Modal Sandbox lifecycle and keep ownership explicit.
- [Automate a browser](https://modal-computer-use.mintlify.app/v1/build/browser-automation.md): Configure Chromium or Firefox in a Modal desktop and open a page.
- [Run action batches](https://modal-computer-use.mintlify.app/v1/build/action-batches.md): Validate and execute ordered desktop actions in one request.
- [Capture screenshots and recordings](https://modal-computer-use.mintlify.app/v1/build/screenshots-recordings.md): Capture desktop images and record an MP4 file.
- [Store and persist artifacts](https://modal-computer-use.mintlify.app/v1/build/artifacts-storage.md): Read, write, download, and persist files under the artifact root.
- [Hand a desktop to a Modal Function](https://modal-computer-use.mintlify.app/v1/build/session-handoff.md): Borrow an existing desktop for one deployed Modal Function trajectory.
- [Run an OpenAI computer loop](https://modal-computer-use.mintlify.app/v1/integrate/openai.md): Connect the OpenAI Responses API computer tool to a Modal desktop.
- [Run an Anthropic computer loop](https://modal-computer-use.mintlify.app/v1/integrate/anthropic.md): Connect the Anthropic Messages API computer tool to a Modal desktop.
- [Deploy a Modal desktop](https://modal-computer-use.mintlify.app/v1/operate/deploy.md): Create, attach, hand off, and clean up Modal desktop Sandboxes.
- [Configure a desktop](https://modal-computer-use.mintlify.app/v1/operate/configure.md): Set desktop, runtime, network, storage, browser, action, and budget policy.
- [Keep warm desktop capacity](https://modal-computer-use.mintlify.app/v1/operate/warm-capacity.md): Fill and claim a bounded pool of ready Modal desktop Sandboxes.
- [Reduce desktop loop latency](https://modal-computer-use.mintlify.app/v1/operate/performance.md): Measure startup and warm operations, then reduce the work on the measured path.
- [Secure a desktop deployment](https://modal-computer-use.mintlify.app/v1/operate/security.md): Protect daemon access, noVNC, provider loops, logs, and retained artifacts.
- [Troubleshoot a desktop](https://modal-computer-use.mintlify.app/v1/operate/troubleshooting.md): Diagnose readiness, X11 input, screenshots, adapters, storage, and warm capacity.
- [Benchmark overview](https://modal-computer-use.mintlify.app/v1/benchmarks/overview.md): Read modal-computer-use benchmark results with their measurement boundaries and evidence status.
- [Current benchmark results](https://modal-computer-use.mintlify.app/v1/benchmarks/current-results.md): View the current July 30 warm-operation benchmark results and their comparison limits.
- [Input capacity in v1](https://modal-computer-use.mintlify.app/v1/benchmarks/input-capacity.md): Keep the v1.1.0 input limit separate from the v2 weighted-capacity result.
- [Latency evidence](https://modal-computer-use.mintlify.app/v1/benchmarks/latency-evidence.md): Review the focused evidence for batching, subprocess execution, native X11 input, and the recorded cost estimate.
- [Run benchmarks](https://modal-computer-use.mintlify.app/v1/benchmarks/run.md): Run credential-free and Modal benchmarks with the repository benchmark commands.
- [Reproduce a benchmark](https://modal-computer-use.mintlify.app/v1/benchmarks/reproducibility.md): Follow the canonical procedure and verify tracked Modal Computer Use benchmark evidence.
- [API reference overview](https://modal-computer-use.mintlify.app/v1/reference/overview.md): Choose the modal-computer-use v1.1.0 Python SDK or daemon API surface.
- [Migrate from v1 to v2](https://modal-computer-use.mintlify.app/v1/reference/migration-v2.md): Move a v1.1.0 application to Modal Computer Use v2.0.1.
- [Configuration reference](https://modal-computer-use.mintlify.app/v1/reference/configuration.md): Configure Modal orchestration with ComputerConfig and configure a direct daemon with environment variables.
- [Namespace reference](https://modal-computer-use.mintlify.app/v1/reference/namespaces.md): Find the v1.1.0 Python methods for mouse, keyboard, screenshots, actions, artifacts, and other daemon capabilities.
- [Errors and models](https://modal-computer-use.mintlify.app/v1/reference/errors-and-models.md): Review stable v1.1.0 result models and top-level error types.
- [OpenAPI reference](https://modal-computer-use.mintlify.app/v1/reference/openapi.md): Use and verify the checked-in OpenAPI 3.1 schema for the v1.1.0 daemon.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.