@oya-ai/cli weekly downloads statistics

@oya-ai/cli

The Command-Line Interface for the Oya Browser Control Plane.

Manage, navigate, benchmark, and watch browser agent instances on Oya Cloud, Browserbase, Steel, Anchor, Browser Use, or your self-hosted Chrome fleet straight from your terminal.

Oya does for browser vendors what OpenRouter does for LLM providers: one interface, and the vendor behind it is a setting.

⚡ 60-Second Setup

🚀 Everyday Workflow

Scripting: --json and exit codes

• --json puts exactly one JSON document on stdout. Prompts and progress go to stderr. • A failure prints one line on stderr (✗ message, then a hint), or with --json one object: {"error", "code", "status"}. • Exit codes: 0 done, 1 the request failed or got no answer, 2 the command could not be run as typed and nothing was sent. • An unknown flag is an error, not ignored. --flag=value works as well as --flag value. • --debug (or OYA_DEBUG=1) adds the status, the response body and the stack to a failure. • oya install, oya init and oya stealth-test are for people and refuse --json.

📖 Command Reference

Fleet Execution & Control

| Command | Flags | Description |

| oya start | [--persona ] [--provider ] [--name ] [--governed] [--budget-usd ] | Launch browser instance and print its ID and CDP endpoint | | oya goto | [--id ] | Navigate to URL (defaults to newest browser) | | oya ask "" | [--id ] | Drive page using configured AI model | | oya open | [--id ] | Launch interactive SSE live view in system browser | | oya ls | [--json] | List active fleet browsers with health status | | oya status | [--id ] [--json] | Detailed metrics, error counts, and recent activity log | | oya rm ... | [--all] | Terminate target browser or entire fleet |

Persona Management (Anti-Ban Identities)

Moving logins (oya cookies)

A persona's cookie jar is its logins. Sign in once, then carry the session to wherever it is needed:

Human-in-the-Loop Takeover

When automation encounters hardware 2FA, phone biometric approvals, or complex verification:

Governance, Auditing & Webhooks

Stealth Benchmarking

Benchmark your browser deployment against live detection platforms (CreepJS and Bot.Sannysoft). The harness lives in server/, so run this from a checkout of the repo, or set OYA_SERVER_DIR:

🤖 CI/CD & Headless Environments

In automated pipelines (GitHub Actions, GitLab CI, Docker), use flags or environment variables to bypass interactive prompts:

| Global Flag | Description |

| --key | Override API key for command | | --url | Override control plane URL | | --id | Target specific browser instance ID | | --json | Output pure JSON for programmatic parsing |

📄 License

MIT © Oya