# ServiceNow/BrowserGym

🌎💪 BrowserGym, a Gym environment for web task automation

Repository: https://github.com/ServiceNow/BrowserGym
Canonical: https://ross.abutalabs.com/products/browsergym
Language: Python
License: NOASSERTION
License Family: other
Topics: ai, llm, webagent
Last push: 2026-07-17T18:09:30+00:00

## Health v2 (maintenance only)
Score: 78/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 93, release rhythm 66, longevity 67
- inputs: {"age_days": 938, "days_push": 47, "days_rel": 225, "gap_med": 3, "n_releases_24m": 28}
- flags: no_license
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1329, forks 192 (observed 2026-08-28T04:04:23.776342+00:00)

## What it is
BrowserGym is a Gym-style environment library for web task automation, providing a unified interface to web agent benchmarks like MiniWoB, WebArena, and WorkArena. It enables researchers to build, test, and evaluate LLM-powered web agents against standardized browser tasks.

## Use cases
- evaluate llm web agents on browser benchmarks
- build a web automation agent with gym environment
- benchmark my agent on webarena and miniwob
- create a custom web task benchmark
- train agents to automate browser tasks

## When to choose
- you are researching or benchmarking LLM-based web agents
- you need a unified Gym API across multiple web benchmarks
- you want to design custom web automation tasks

## When to avoid
- you need a production browser automation tool for end users
- you want simple scraping without an agent/research framework
- you need a consumer product - it is explicitly research-oriented

## Facets
- artifact type: library
- maturity: active
- function: agent-framework, machine-learning, llm-inference, testing, benchmarking
- domain: artificial-intelligence, large-language-models, web-development
- platform: python, cross-platform
- tags: web-agent, gym-environment, browser-automation, reinforcement-learning-environment, web-benchmarks, llm-agents, ai-agents, research

## Member repositories
- ServiceNow/BrowserGym (main) score 78

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:04:23.776342+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 2026-08-30T04:45:37.753558+00:00, confidence not recorded.
  - readme: https://github.com/ServiceNow/BrowserGym (fetched 2026-08-28T04:04:23.776342+00:00, sha 577c7ecf7809)
- Data as of 2026-08-30T08:39:29.467469+00:00.