# openai/openai-cua-sample-app

Learn how to use CUA (our Computer Using Agent) via the API on multiple computer environments.

Repository: https://github.com/openai/openai-cua-sample-app
Canonical: https://ross.abutalabs.com/products/openai-cua-sample-app
Language: TypeScript
License: MIT
License Family: permissive
Last push: 2026-03-30T16:48:59+00:00

## Health v2 (maintenance only)
Score: 53/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 74, release rhythm 35, longevity 38
- inputs: {"age_days": 540, "days_push": 156, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1775, forks 427 (observed 2026-08-28T04:05:34.492140+00:00)

## What it is
A TypeScript sample application from OpenAI demonstrating how to use the Computer Using Agent (CUA) via the Responses API against browser environments. It includes a Next.js operator console, a Fastify runner managing browser sessions and workspaces, and shared packages for defining and verifying automation scenarios.

## Use cases
- learn how to integrate OpenAI computer-use agents via the API
- automate browser tasks with an AI agent clicking and typing
- build an operator console for reviewing agent runs and screenshots
- compare native computer-use mode versus scripted Playwright code mode
- set up isolated browser workspaces for AI agent scenarios
- replay and inspect agent event traces and screenshots

## When to choose
- you want a reference implementation for OpenAI's CUA / Responses API
- you need to prototype browser automation driven by an LLM agent
- you want a working example of an operator UI with run replay and event logs
- you are evaluating native computer-use actions versus code-execution browser control

## When to avoid
- you need a production-ready agent framework rather than a sample app
- you want to automate desktop applications instead of browser sessions
- you need Python examples - the legacy Python sample is not in this branch
- you cannot provide an OpenAI API key, since live runs are secret-gated

## Facets
- artifact type: learning-resource
- maturity: active
- function: agent-framework, llm-inference, web-scraping, http-server, developer-tools
- domain: artificial-intelligence, large-language-models, developer-tools, web-development
- platform: cross-platform
- tags: computer-use-agent, cua, openai-api, sample-app, browser-automation, playwright, responses-api, operator-console, nextjs, fastify, ai-agents, automation, nodejs, typescript, docker

## Member repositories
- openai/openai-cua-sample-app (main) score 53

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:05:34.492140+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T03:25:21.692651+00:00, confidence not recorded.
  - readme: https://github.com/openai/openai-cua-sample-app (fetched 2026-08-28T04:05:34.492140+00:00, sha 8deb8a6dbe4e)
- Data as of 2026-08-30T08:39:29.467469+00:00.