{
  "aliases": [],
  "authors": [],
  "details": {
    "benchmark_protocol": {
      "actions": [
        "element-ID actions",
        "coordinate mouse and keyboard actions",
        "tab/navigation/scroll operations",
        "Python/Playwright action mode"
      ],
      "counts": [],
      "evaluation": "Gymnasium reset/step interface delegates reward, completion, messages and task info to each adapter’s validate() method. There is no universal BrowserGym success metric or shared task total; retain benchmark/adapter version and action subset for comparisons.",
      "evidence_urls": [
        "https://raw.githubusercontent.com/ServiceNow/BrowserGym/HEAD/README.md",
        "https://raw.githubusercontent.com/ServiceNow/BrowserGym/HEAD/browsergym/core/src/browsergym/core/env.py",
        "https://raw.githubusercontent.com/ServiceNow/BrowserGym/HEAD/docs/src/core/action_space.rst",
        "https://api.github.com/repos/ServiceNow/BrowserGym/git/trees/main?recursive=1"
      ],
      "focus": [
        "shared browser-agent evaluation interface",
        "benchmark adapters with separate task definitions"
      ],
      "notes": "A framework with interactive browser adapters and a static WebLINX evaluation integration, so the profile setting is mixed. README-listed integrations are not added as disjoint tasks or datasets. Core requires Playwright/Chromium; individual benchmarks have their own setup, credentials, licenses and validation rules. Apache-2.0 covers repository code only. No browser tasks or third-party project code were executed.",
      "observations": [
        "browser screenshot",
        "DOM snapshot",
        "accessibility tree",
        "tab/URL state",
        "previous action and its error",
        "chat messages and task goal"
      ],
      "setting": "mixed"
    }
  },
  "environments": [
    "web"
  ],
  "evidence": [
    {
      "method": "editorial_primary_review",
      "retrieved_at": "2026-10-16T06:35:57Z",
      "role": "primary",
      "sha256": "7b685115ef822989fac9093518d16c46f0202736d005eed56e7f9fc36b43f3ee",
      "supports": [
        "name",
        "kind",
        "summary",
        "environments",
        "topics",
        "scope",
        "relations",
        "details.benchmark_protocol"
      ],
      "url": "https://raw.githubusercontent.com/ServiceNow/BrowserGym/HEAD/README.md"
    },
    {
      "method": "editorial_primary_review",
      "retrieved_at": "2026-10-17T14:22:05Z",
      "role": "primary",
      "sha256": "867cc6efdfead356839fef8a26b492502b8b5e3d165d1b3c17d24aaeec268d63",
      "supports": [
        "details.benchmark_protocol"
      ],
      "url": "https://raw.githubusercontent.com/ServiceNow/BrowserGym/HEAD/browsergym/core/src/browsergym/core/env.py"
    },
    {
      "method": "editorial_primary_review",
      "retrieved_at": "2026-10-17T14:21:11Z",
      "role": "primary",
      "sha256": "c4bbe6b7d0cb36a5d5effab1aa5568301cead97fe4015273e1b82c667da3823a",
      "supports": [
        "details.benchmark_protocol"
      ],
      "url": "https://raw.githubusercontent.com/ServiceNow/BrowserGym/HEAD/docs/src/core/action_space.rst"
    },
    {
      "method": "editorial_primary_review",
      "retrieved_at": "2026-10-17T14:14:08Z",
      "role": "primary",
      "sha256": "418efbe8c65ad7ebab61d3a066339876c820a9c957506eb456b8078ab352f6bf",
      "supports": [
        "details.benchmark_protocol"
      ],
      "url": "https://api.github.com/repos/ServiceNow/BrowserGym/git/trees/main?recursive=1"
    },
    {
      "method": "license_header",
      "retrieved_at": "2026-10-16T06:41:57Z",
      "role": "primary",
      "sha256": "b192c58991e8ff585cc574615d40e74185404d4b96c1109d423071ab1367344b",
      "supports": [
        "licenses.code"
      ],
      "url": "https://raw.githubusercontent.com/ServiceNow/BrowserGym/HEAD/LICENSE"
    }
  ],
  "id": "github:servicenow/browsergym",
  "identifiers": {},
  "kind": "benchmark",
  "last_verified_at": "2026-10-16T06:35:57Z",
  "licenses": {
    "code": {
      "basis": "repository LICENSE header; not a weights/data license",
      "declared": "Apache-2.0",
      "retrieved_at": "2026-10-16T06:41:57Z",
      "url": "https://raw.githubusercontent.com/ServiceNow/BrowserGym/HEAD/LICENSE"
    },
    "data": null,
    "document": null,
    "weights": null
  },
  "links": [
    {
      "rel": "primary",
      "url": "https://github.com/ServiceNow/BrowserGym"
    }
  ],
  "name": "BrowserGym",
  "notes": "",
  "published_at": null,
  "relations": [
    {
      "evidence_url": "https://raw.githubusercontent.com/ServiceNow/BrowserGym/HEAD/README.md",
      "target_id": "arxiv:2412.05467",
      "type": "documented_by"
    }
  ],
  "review_status": "curated",
  "scope": "core",
  "summary": "Unified interfaces for browser-agent tasks and benchmark environments.",
  "topics": [
    "evaluation"
  ],
  "updated_at": null,
  "url": "https://github.com/ServiceNow/BrowserGym"
}