Skip to main content
A Computer-Use Agent sees the screen and decides what to click, type, and scroll. Use it when the work lives behind a user interface with no API to call. See it as a visual agent that directly interprets and acts on the UI. The Computer-Use Agent API gives you programmatic control over agents built on top of H’s Holo family of Vision Language Models. You describe a task in plain language; H provisions the environment, runs the agent, and returns the result through a lifecycle your app can monitor, steer, and stop along the way. V1 ships with the Browser environment; Desktop is in What’s next. One call to start a session: 
curl -X POST https://agp.eu.hcompany.ai/api/v2/sessions \
  -H "Authorization: Bearer $H_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"agent": "h/web-surfer-holo3-1-35b", "messages": [{"type": "user_message", "message": "On Google Flights, find the cheapest direct flight from Paris (CDG) to Tokyo (NRT) this Saturday. Return the airline and the price."}]}'
That one call spins up the environment, hands it to the built-in h/web-surfer-holo3-1-35b agent, and runs until there’s an answer. The Quickstart takes it from there.

​
How it fits together

The building blocks, each with a dedicated page:
ConceptWhat it is
AgentA reusable configuration: environments, skills, model, and instructions. Reference a pre-built agent or create your own.
EnvironmentThe surface the agent acts on and observes. A browser today; desktop in What’s next.
SessionA single run of an agent against a task, with a lifecycle you can steer.
SkillA reusable instruction fragment an agent loads on demand. Optional.
Beta: this API is still evolving. Send questions or bug reports to feedback@hcompany.ai.

​
Get started

Quickstart

Create your first agent session in under 5 minutes. Start here.

Integrations

Install the Python or TypeScript client and skip the polling boilerplate.

Session lifecycle

See agent, environment, and session come together at runtime.

Agent View

Watch a session live or replay it afterward to debug, monitor, and follow your agents.

​
Integrations

Call the API directly over HTTP, or use a typed client. Both ship as hai-agents — see the Integrations page for install, auth, and the helpers we recommend in production.

Python

Sync and async clients, typed with Pydantic v2.

TypeScript

A fully typed client for sessions, agents, skills, and environments.