crewai-plasmate

Plasmate web browsing tool for CrewAI agents. Get clean semantic page content using 10-16x fewer tokens than Chrome.

Install

pip install crewai-plasmate

Usage

from crewai import Agent, Task, Crew
from crewai_plasmate import PlasmateFetchTool

researcher = Agent(role="Researcher", tools=[PlasmateFetchTool()])
task = Task(description="Fetch https://example.com and summarize it", agent=researcher)
Crew(agents=[researcher], tasks=[task]).kickoff()

What is Plasmate?

Plasmate is a lightweight web browsing engine for AI agents. It returns a Semantic Object Model (SOM) instead of raw HTML, reducing token usage by 10-16x compared to headless Chrome.

Links

• Plasmate documentation
• SOM/1.0 specification
• Public benchmark
• W3C Web Content for AI Community Group
• Plasmate engine on GitHub

Publishing

This package publishes to PyPI via OIDC trusted publisher on every v* tag push (see .github/workflows/publish.yml).

To publish a new version:

1. Bump version in pyproject.toml
2. git tag v<version> && git push --tags
3. Workflow runs; package appears on PyPI

First-time PyPI setup (one-time, owner-only):

• Create the project on PyPI by visiting https://pypi.org/manage/account/publishing/
• Add a trusted publisher with: project=crewai-plasmate, owner=plasmate-labs, repo=crewai-plasmate, workflow=publish.yml, environment=pypi

License

Apache 2.0

---

Part of the Plasmate Ecosystem

Engine	plasmate - The browser engine for agents
MCP	plasmate-mcp - Claude Code, Cursor, Windsurf
Extension	plasmate-extension - Chrome cookie export
SDKs	Python / Node.js / Go / Rust
Frameworks	LangChain / CrewAI / AutoGen / Smolagents
Tools	Scrapy / Audit / A11y / GitHub Action
Resources	Awesome Plasmate / Notebooks / Benchmarks
Docs	docs.plasmate.app
W3C	Web Content Browser for AI Agents