# web-arena-x/webarena

Code repo for "WebArena: A Realistic Web Environment for Building Autonomous Agents"

Repository: https://github.com/web-arena-x/webarena
Canonical: https://ross.abutalabs.com/products/webarena
Homepage: https://webarena.dev
Language: Python
License: Apache-2.0
License Family: permissive
Topics: agent, nlp
Last push: 2025-11-26T21:16:00+00:00

## Health v2 (maintenance only)
Score: 43/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 54, release rhythm 8, longevity 81
- inputs: {"age_days": 1136, "days_push": 280, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1588, forks 246 (observed 2026-08-28T04:05:08.054630+00:00)

## What it is
WebArena is a standalone, self-hostable web environment and benchmark for building and evaluating autonomous web agents. It provides realistic websites (shopping, forums, GitLab, maps) with ~170 tasks, execution trajectories, and tooling to reproduce benchmark results from the NeurIPS 2024 paper.

## Use cases
- benchmark my web navigation agent on realistic tasks
- self-host a sandboxed web environment for testing autonomous agents
- evaluate LLM agents on multi-step web tasks
- compare my agent against the WebArena leaderboard
- get human demonstration trajectories for training web agents
- reproduce results from the WebArena paper

## When to choose
- you need a reproducible, self-hosted benchmark for autonomous web agents
- you want realistic multi-step web tasks rather than static datasets
- you want to compare against published WebArena leaderboard results

## When to avoid
- you need large-scale parallel evaluation with modern infrastructure - the maintainers recommend AgentLab/BrowserGym instead
- you want a simple hosted API without self-hosting websites
- you need multimodal/visual tasks specifically - consider VisualWebArena

## Facets
- artifact type: framework
- maturity: active
- function: agent-framework, benchmarking, testing, web-scraping
- domain: artificial-intelligence, large-language-models, developer-tools
- platform: python, self-hosted
- tags: web-agents, benchmark, autonomous-agents, llm-agents, web-navigation, evaluation, self-hostable, ai-agents, research, docker, linux, macos

## Member repositories
- web-arena-x/webarena (main) score 43

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:05:08.054630+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 2026-08-30T03:55:05.083413+00:00, confidence not recorded.
  - readme: https://github.com/web-arena-x/webarena (fetched 2026-08-28T04:05:08.054630+00:00, sha 59f8e2bf1bd9)
  - homepage: https://webarena.dev (fetched 2026-08-29T11:25:51.555990+00:00, sha a09fc6022aa8)
- Data as of 2026-08-30T08:39:29.467469+00:00.