WootzappLab/browser-agent-tasks

1,037 downloads in 30 days0 likesLast modified 30 Oct 2026View on Hugging Face

The declared licence (apache-2.0) is classified as permissive.

Declared licence: apache-2.0, from the tagEvidence: README.md at 7c961cc
Licence files: No licence file at the top level of the repositoryEvidence: file list at 7c961cc
Declared sources: None listed in the card metadata
Access gated on Hugging Face: No

Each cell links to what was checked: a file, a file list or a Hub record.

Citation text from the card

WebVoyager introduced multimodal automatic evaluation for open-ended web
tasks [[1]](#references). The screenshot baseline follows the Universal
Verifier's rubric-centered design: criterion-specific evidence selection,
separate process and outcome judgments, and failure attribution
[[2]](#references). WootzApp adds the DOM-based adaptation, browser-native
evidence capture, paired comparison protocol, and source-aware audit.

WebArena provides screenshot, HTML DOM, and accessibility-tree observations
and evaluates functional task completion in controlled websites
[[4]](#references). These systems provide context for structured and visual
browser evidence. This benchmark measures what paired recorded
representations preserve under a shared rubric; it does not reproduce those
benchmarks' task-success experiments.

All datasets marked Permissive or share-alike | JSON

The same facts for your agent

A free MCP server at mcp.datalicense.uk/mcp. Connect it from an MCP client such as Claude or Cursor.

dataset_licence({ "id": "WootzappLab/browser-agent-tasks" })
signal: "permissive"
stated_licence: "apache-2.0"
sources: 0