Stagehand Ruby
https://github.com/browserbase/stagehand-ruby- Category
- Developer Tools
- Rank
- No. 1148Tools index
Previous survey · No. 1153 ·
- Pricing
- Open Source
- Type
- TOOL
- Builder
- browserbase
- GitHub
- 9 stars
- Latest release
- v3.21.0
- Date
About
Official Stagehand AI browser automation SDK for Ruby. Built by Browserbase.
What it does
Stagehand Ruby lets Ruby applications create browser sessions, navigate pages, issue natural-language actions, extract structured data, and run multi-step agents. It talks to a hosted or locally managed server and can stream progress through Server-Sent Events.
Why it's ranked here
The library covers the full session workflow with typed request and response models, streaming variants, configurable retries, specific error classes, and local execution support. Its generated foundation is practical, while the custom local server manager adds meaningful capability beyond a thin REST wrapper.
What's good
It ships Yard documentation plus RBS and RBI types. Client instances are thread-safe and use pooled HTTP connections. Retry behavior covers network failures, timeouts, rate limits, conflicts, and server errors. Callers can also pass undocumented fields or reach undocumented endpoints without losing authentication and retry handling.
Tradeoffs
Remote use requires Browserbase and model provider credentials. Local mode may download and launch a separate platform-specific server binary. Direct page control in the documented workflows adds Playwright or Watir and their browser dependencies. Timeouts are retried by default, so poorly chosen settings can extend failure latency.
How to use it well
Choose it for Ruby services that need natural-language browser actions, schema-guided extraction, or autonomous multi-step browsing through Stagehand. Keep one client per application to benefit from connection pooling. Add Playwright or Watir when you need direct page-level browser control. It does not replace those browser-control libraries.
Technical notes+
README.md requires Ruby 3.2.0+ and documents Bundler installation from the Git repository. lib/stagehand/client.rb defines Stagehand::Client with remote and local server modes, a 60-second default timeout, two default retries, and session resources. lib/stagehand/resources/sessions.rb exposes synchronous and SSE streaming operations for actions, extraction, observation, execution, navigation, replay, session start, and termination. lib/stagehand/internal/stream.rb parses SSE lifecycle events and converts them into typed models. lib/stagehand/local.rb selects macOS, Windows, or Linux binaries for arm64 or x64, downloads releases, caches the executable, starts it on a free local port, probes readiness endpoints, and terminates the process during cleanup. lib/stagehand/errors.rb maps HTTP and transport failures to specific exception classes.
Observed
- License
- MIT License
- Primary language
- Ruby
- Runtime
- Ruby 3.2.0 or newer
- Packaging and installation
- Stagehand gem installed through Bundler from the Git repository
- Interface
- Ruby library client for the Stagehand REST API, with Server-Sent Events streaming
- Execution modes
- Hosted remote server or locally managed server binary
- Local platform support
- macOS, Windows, and Linux binaries for arm64 and x64
Read from README.md, Gemfile, lib/stagehand.rb, lib/stagehand/local.rb, lib/stagehand/client.rb, lib/stagehand/errors.rb, lib/stagehand/models.rb, lib/stagehand/version.rb, lib/stagehand/internal.rb, lib/stagehand/file_part.rb, lib/stagehand/request_options.rb, lib/stagehand/internal/util.rb, lib/stagehand/models/action.rb, lib/stagehand/internal/stream.rb, lib/stagehand/resources/sessions.rb.
What it can do
Automate browser interactions using AI
Natural language instructions and target web page → Executed browser actions (clicks, form fills, navigation)
Extract data from web pages
Web page URL and data extraction requirements → Structured data from the webpage
Fill out web forms automatically
Form data and target form on webpage → Completed and submitted web form
Navigate between web pages
Navigation instructions and target URLs → Browser positioned at desired webpage
Click on web elements intelligently
Element description and target webpage → Clicked web element (buttons, links, etc.)
Take screenshots of web pages
Target webpage or browser state → Screenshot image file
Wait for page elements to load
Element selector or loading condition → Confirmation of element availability
Tags
Tech Stack
Media
Comments (0)
No comments yet
Editorially curated, with community endorsements as a secondary signal. Corrections welcome.