# McGill-NLP/webllama

Llama-3 agents that can browse the web by following instructions and talking to you

Repository: https://github.com/McGill-NLP/webllama
Canonical: https://ross.abutalabs.com/products/webllama
Homepage: https://webllama.github.io
Language: Python
License: MIT
License Family: permissive
Last push: 2026-08-16T18:12:12+00:00

## Health v2 (maintenance only)
Score: 59/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 98, release rhythm 8, longevity 61
- inputs: {"age_days": 866, "days_push": 17, "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 1402, forks 102 (observed 2026-08-28T04:04:37.697502+00:00)

## What it is
WebLlama is a framework for building Llama-3-powered agents that browse the web by following natural language instructions and dialogue. It includes training scripts, evaluation benchmarks (WebLINX), datasets, and deployment integrations with Playwright, Selenium, and BrowserGym.

## Use cases
- build an agent that browses the web for me
- train an LLM on web navigation tasks
- evaluate a model on real-world website navigation
- fine-tune Llama 3 on browser interaction data
- automate web browsing via dialogue instructions
- integrate LLM agents with Playwright or Selenium

## When to choose
- you want to build or fine-tune LLM agents for web navigation
- you need the WebLINX benchmark or curated web interaction datasets
- you want a research-grade starting point for dialogue-driven browsing agents

## When to avoid
- you need a production-ready, turnkey browser automation tool
- you only need simple scraping without an LLM agent
- you cannot run or fine-tune 8B-scale models

## Facets
- artifact type: framework
- maturity: active
- function: agent-framework, llm-training, machine-learning, web-scraping, chatbot
- domain: large-language-models, web-development, machine-learning
- platform: python, cross-platform
- tags: llama-3, web-navigation, weblinx, browser-automation, dialogue-agents, playwright, selenium, ai-agents, natural-language-processing

## Member repositories
- McGill-NLP/webllama (main) score 59

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:04:37.697502+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 2026-08-30T04:38:57.363383+00:00, confidence not recorded.
  - readme: https://github.com/McGill-NLP/webllama (fetched 2026-08-28T04:04:37.697502+00:00, sha f600f798b0cd)
  - homepage: https://webllama.github.io (fetched 2026-08-29T11:53:07.038572+00:00, sha ad839a86eaf2)
  - registry_pypi: https://pypi.org/pypi/webllama/json (fetched 2026-08-29T11:53:07.048811+00:00, sha 990ff25759d7)
- Data as of 2026-08-30T08:39:29.467469+00:00.