Practical examples showing common Airbrowser workflows.
# Start Airbrowser
docker compose up --build
# Install the Python client
uv pip install -e generated-clients/python| File | Description |
|---|---|
basic_navigation.py |
Create browser, navigate, screenshot, cleanup |
form_automation.py |
Fill forms, click buttons, handle inputs |
mcp_agent_example.py |
MCP integration for AI agents |
parallel_browsers.py |
Run multiple browsers concurrently |
proxy_rotation.py |
Different proxy per browser instance |
what_is_visible.py |
AI vision analysis of page content |
cloudflare_captcha_vision.py |
AI vision to detect and click captcha |
# Basic navigation
uv run python examples/basic_navigation.py
# Form automation
uv run python examples/form_automation.py
# MCP agent (requires: uv pip install fastmcp)
uv run python examples/mcp_agent_example.py
# Parallel browsers
uv run python examples/parallel_browsers.py
# Proxy rotation (edit PROXIES list first)
uv run python examples/proxy_rotation.pyThese examples require VISION_API_BASE_URL, VISION_API_KEY, and VISION_MODEL in the container:
# Start with vision enabled
VISION_API_BASE_URL=https://your-openai-compatible-endpoint/v1 \
VISION_API_KEY=your-api-key \
VISION_MODEL=your-vision-model \
docker compose up
# AI vision analysis
uv run python examples/what_is_visible.py
# Cloudflare captcha detection and click
uv run python examples/cloudflare_captcha_vision.pyAll services at http://localhost:18080:
- Dashboard: http://localhost:18080/
- REST API Docs: http://localhost:18080/docs/
- MCP Endpoint: http://localhost:18080/mcp
- VNC Viewer: http://localhost:18080/vnc/