RoboRuby
Posts
Ruby AI News - June 10th, 2025

Ruby AI News - June 10th, 2025

A ton of Ruby AI alpha drops as the community ramps up

Matt Solt
June 10, 2025

Welcome to the 8th Edition of the Ruby AI Newsletter! This edition features Kieran Klaassen dropping a ton of Ruby AI alpha, the production-tested Raix AI library hitting 1.0, San Francisco positioning itself to be the center of the Ruby AI universe, and much more.

Top Stories
Content
Events
- Previous
- Upcoming
Open Source Updates
Jobs & Opportunities

Content

The top Ruby AI content from across the web and social media.

Announcement: The Model Context Protocol organization released the the official Ruby SDK for the model context protocol, an open protocol that standardizes how applications provide context to LLMs. The software development kit is maintained in collaboration with Shopify, and can be installed with the mcp gem.

Update: Carmine Paolino shipped version 1.3.0 of RubyLLM. The update adds auto-detection of files in chat, isolated configuration contexts for multi-tenant applications, support for Ollama and OpenRouter, and enhanced Rails integration with ActiveStorage. Full documentation and guides are available at rubyllm.com.

Podcast: Valentino Stoll and Joe Leo officially launched The Ruby AI Podcast. The first episode featured a discussion on phoenix.love, the top Ruby AI libraries, and the importance of properly evaluating AI tools. The second episode included guest Scott Werner for a talk about Sublayer, the ArtificialRuby meetup, and the Ruby community at large. The third episode hosted guest Justin Bowen for a deep dive on ActiveAgent, a Rails-native AI toolkit. The fourth episode, coming June 18th, will feature Obie Fernandez and will look at Roast, the open-source framework for building structured AI workflows with Ruby.

Tutorial: Adrien Payong and Shaoni Mukherjee of DigitalOcean published How to Build AI Agents with Ruby. The article explores setting up the ruby-openai gem, defining an agent, creating an interactive loop, and extending the agent with tools and advanced memory.

Update: Abdelkader Boudih released version 0.54 of action_mcp, a specification-compliant Rails engine for model context protocol servers. The release features major new updates including a swappable session store to persist sessions across different environments, a gateway authentication system, and implementation of the Oauth 2.1 spec.

Launch: Abdelkader Boudih also launched ContribOSS (COSS). The COSS standard is intended to provide universal access to open source components, prevent vendor lock-in, ensure license and ecosystem compatibility, and provide an ethical framework for AI projects. COSS projects would include standardized metadata in a coss.toml file to help AI tools understand project structure, build processes, and conventions, reducing computational waste and enabling more accurate AI assistance.

Tutorial: Conner Jensen recorded a tutorial on using Cursor with Taskmaster entitled Code Rails 10x Faster: AI-Powered Dev with Cursor. He also detailed his learnings, what worked well, and future explorations in this Reddit post.

Tips: Conner Jensen also shared his AI-enabled workflow for building Rails applications with this Tweet and video.

Article: Paweł Strzałkowski published MCP Client in Rails using ruby-mcp-client gem, walking through implementing an LLM-driven assistant with Rails, ruby-openai, and the ruby-mcp-client gem. This is a follow up to the article MCP Server with Rails and fast-mcp.

Tutorial: Cezar Halmagean recorded a video on How to Build an AI Sales Agent With Ruby on Rails. The video walks through the Rails application structure, schema, and integration with the ruby-openai and langchainrb gems. The source code for the app is available on Github.

Announcement: Valentino Stoll released Agentic, a Ruby gem for building and running AI agents in a plan-and-execute fashion. Agentic provides a CLI and library to build, manage, and deploy asynchronous AI agents using OpenAI's LLM API. A demo video is available, and the source code is up on Github.

Announcement: Valention Stoll also open-sourced AI Software Architect, a markdown-based framework for implementing rigorous software architecture practices, with specialized AI assistant integrations. The component video shares the structured approach to documentation, reviews, and progress tracking when working with AI coding assistants.

Article: Kevin Sylvestre wrote about Exploring Common AI Patterns with Ruby. The article explores three integration patterns leveraging OmniAI, a gem to standardize APIs for multiple AI/LLM providers. Examples include parsing PDF receipts into a CSV file, retrieval-augmented generation for searching product manuals, and building an AI web browsing agent.

Update: Yorick Jacquin announced the release of fast-mcp version 1.5.0. Fast MCP is a ruby implementation of the model context protocol. The update brings filtering for tools and resources, support for resource templates, and authorization and request headers for tool calls.

Announcement: Alex Larra created Internator, an AI-powered pull request sidekick. Internator is a Ruby CLI that sends a diff to OpenAI Codex, applies a single incremental change per iteration, auto-commits with an AI-crafted message, and repeats until your objectives are met.

Update: Mario Alberto Chávez released a major update for the Rails MCP Server. Version 1.2.0 brings complete Rails guides, Turbo and Stimulus documentation, and Kamal deployment guides into your AI conversations. Read all about it in Rails MCP Server: Enhanced Documentation Access and AI Workflow Integration.

Announcement: Radamés Roriz shared benchmark data for ActiveGenie in this tweet:

The latest run used exactly 2,101,930 tokens, across 1,358 requests, taking 1h 38min on GitHub Actions. 😱

That's the level of effort we're putting in to ensure consistency and reliability across LLMs.

https://x.com/radamesroriz/status/1930029135075926286

Demo: As part of the Experimental AI series “Works on My Machine”, Scott Werner shared a video and article on LLM-powered Method Resolution with Synonllm. Synonllm is a Ruby module that combines method_missing with an LLM to provide AI-powered synonym matching for method calls.

Podcast: Scott Werner was also a guest on Dead Code for the episode Wish Granted. The episode discusses Monkey’s Paw, a Ruby-based web framework that uses natural language prompts to generate web pages, embracing LLM hallucinations as a creative feature.

Article: Mateusz Palak detailed building ActiveRubyist over the past year in this Tweet and in the article What I learned while building ActiveRubyist.

Launch: Brian Casel announced new AI components, including OpenAI and Anthropic integration, for Instrumental Components. Instrumental Components is a UI component library and Rails starter kit. There is a companion video for the launch that explains the components, demos an example Rails integration, and includes a section on how to manage AI costs for your SaaS application.

Article: Amanda Bizzinotto of OmbuLabs shared how to build an LLM-based AI Assistant for the FastRuby.io Newsletter. The AI assistant curates articles and generates summaries for the newsletter using Sinatra, pgvector, and Langchain.rb. Amanda also explained Implementing Semantic Search with Sequel and pgvector using Sequel in a follow-up article.

Tutorial: Miles Georgi published Creating an AI Agent with the foobara-agent-cli Ruby Gem, documenting how to build an AI agent with Foobara, a gem for encapsulating domain operations as commands.

Demo: Prabin Poudel shared an example app to answer questions about his blog. The application is a Gemini AI bot built with Rails and RubyLLM. The source code is available on Github.

Article: Valentin Colato of Sinaptia discussed The Untold Challenges of OpenAI’s Batch Processing API using a reliable polling system to handle token limits, partial failures, and retries.

Announcement: Junie, an AI coding agent from JetBrains, is now available in the RubyMine integrated development environment. I was unable to independently verify their claim of solving 60.8% of tasks on a single run on the SWE-bench verified dataset, however that result would only put Junie at 22nd on the leaderboard. You can read more about the latest RubyMine developments in What’s Next for RubyMine.

Article: Giménez Silva Germán Alberto shared an architecture for Automating Document Generation with AI in Ruby on Rails. The code samples use ruby-openai to generate product requirement documents in different file formats.

Opinion: Thoughtbot recorded AI in Focus: Where AI fits in the TDD workflow. The video explores where AI best fits into the test driven development workflow while building a feature for a Rails application. Other Rails and AI content include a recap of the prior month’s video, Building a Rails Feature with AI and Active Job. Thoughtbot also posted videos on Can we recreate a Rails page in JavaScript with AI? and The real AI adoption challenge: your team. Additionally, they added the article AI meets AgeTech: Where predictive care and digital companions are headed to their AI content, and published the final article in their AI for Business series, Implementation strategy.

Article: Irina Nazarova of Evil Martians wrote about A Tea Break: building sfruby.com with Bolt.new. The write up looked at using bolt.new, an AI chatbot for building websites and applications, to quickly create the website for the San Francisco Ruby Conference.

Announcement: José Valim outlined the general vision for Tidewave.ai, a collection of tools that speed up development with AI agents by understanding your web application, in Announcing Tidewave: beyond code intelligence. This coincides with the release of Tidewave for Rails version 0.1.3.

Tips: Pete Hawkins provided a sample Github issue you can run through the Microsoft Copilot agent every day to improve your Rails test coverage in this Tweet. He also created a thread detailing everything he learned during the process. Additionally, he published his working setup for reasoning, planning, and writing code changes in a Github Gist.

Opinion: Noel Rappin noted concerns and considerations on working with AI in What Do I Think I Think About LLMs.

Article: The Rails Drop asked what is Cursor AI? and explained Why It’s Changing the Way We Code in 2025.

Opinion: Justin Searls looked at Why Agents are Bad Pair Programmers because they can code faster thank humans think.

Feedback: Rob Bazinet asked how teams are using AI to help their users in their Ruby on Rails applications. You can view the responses in this X/Twitter thread.

Opinion: David Heinemeier Hansson shared the Ars Technica article OpenAI slams court order to save all ChatGPT logs, including deleted chats on X/Twitter, with the comment:

Don't ask AI anything you wouldn't want opposing counsel to read during discovery. Anything you say may indeed be used against you in a court of law. This is infinitely worse than "your Google searches reveal a lot about you".

https://x.com/dhh/status/1930567527186289057

Events

The San Francisco Ruby Meetup Group met on June 4th, and Irina Nazarova wrote up a recap of the event. In addition to announcements for the Ruby AI Hackathon sign up and the San Francisco Ruby Conference call for proposals, there were four speakers discussing Ruby AI projects:

Andrei Bondarev talked about Langchain.rb and provided an update on version 2.0 which will include a tighter integration with Rails. [Timestamped YouTube link]

Kieran Klaassen introduced Leva, a framework for systematic large language model evaluation with Rails. [Timestamped YouTube link]

Justin Bowen demoed ActiveAgent, a model view controller-type agent framework for AI. [Timestamped YouTube link]

Miles Georgi presented Foobara, an AI-intergrated, command-centric framework for complex domains. [Timestamped YouTube link]

Previously, the San Francisco Meetup Group also hosted Obie Fernandez on May 21st to discuss Structured AI Workflows at Scale with Roast. [Timestamped YouTube link]

Upcoming

June 14th: The Baltic Ruby conference will have a talk by Chris Hasiński on Image Vector Search on June 14th in Riga, Latvia.

June 17th: Lyon.rb will host Guillaume Briday on June 17th in Lyon, France for a conversation on Smarter Rails apps with AI, How to Use LLMs in Real Life.

June 18th: Ruby on Rails Schweiz is meeting on June 18th in Zürich, Switzerland will have a demonstration by Ricardo Carlesso on Rails 8 responsive chat done with RubyLLM.

June 19th: The Brighton Ruby conference will feature a presentation by Valerie Woolard on vector data types and how they are used in Large Language models and AI. The talk on June 19th in Brighton, Great Britian will discuss how LLM models may fall short and demonstrate an example of how to build a similarity search with Ruby.

June 24th: Ruby Europe is hosting a Ruby + AI Meetup on June 24th in Berlin, Germany and will have a multitude of talks on Ruby and AI including Paweł Strzałkowski with the Joy of Creativity in the Age of AI and Chris Hasiński with AI, Offline.

July 16th: ArtificialRuby is hosting a meetup at Betaworks in New York City on July 16th. If you are interested in presenting at an ArtificialRuby meetup, please fill out this form and let them know!

July 19th: The San Francisco Ruby Meetup Group is hosting a Ruby AI Hackathon on July 19th in San Francisco. This one-day hackathon will explore how AI can help solo developers supercharge their Rails application development. Read more about the event from David Doolin’s announcement and Kamil Nicieja’s Tweet.

September 18th: The EuRuKo (European Ruby Conference) on September 18th and 19th in Viana do Castelo, Portugal has announced four! Ruby AI-related speakers:

Obie Fernandez: Roasting Code for Fun & Profit with Structured AI Workflows
Paweł Strzałkowski: Making Rails AI-Native with the Model Context Protocol
Carmine Paolino: RubyLLM: Making AI Development Beautiful Again
Lucian Ghinda: Don't Let Your AI Guess: Teach It to Test!

Open Source Updates

Code Spotlight

Andy Waite open-sourced react_to_rails, an AI-powered tool for converting Tailwind UI React components to Rails View Components. The code uses the openai gem to take your provided React example and return a ViewComponent subclass containing the Ruby component:

react_to_rails/lib/react_to_rails/convert.rb at main · andyw8/react_to_rails

Converts React components to Rails View Components using AI. - andyw8/react_to_rails

github.com/andyw8/react_to_rails/blob/main/lib/react_to_rails/convert.rb

New Versions

Links to release notes or changelog if available.

actionmcp 0.54.0 - Rails Engine with MCP compliant spec.
active_genie 0.26.6 - A set of Ruby modules to work with Generative AI.
aia 0.9.5 - CLI assistant for generative AI workflows to craft, manage, and execute prompts.
boxcars 0.8.3 - Build applications with composability with LLM's.
bundler_mcp 0.2.1 - MCP server that enables AI agents to query information about gems in a Ruby project's Gemfile.
dynamicschema 1.0.0 - Flexible semantic schema definitions for constructing and validating complex configurations.
fast-mcp 1.5.0 - Implementation of the Model Context Protocol.
foobara 0.0.130 - An AI-integrated command-centric framework for complex domains.
git_auto 0.3.0 - AI-powered git commit message generator.
gitingest 0.7.0 - CLI tool that fetches files from a GitHub repository and generates a consolidated text prompt for your LLMs.
highs 0.2.6 - Linear optimization for Ruby.
lammy 0.11.0 - LLM library for Ruby.
llm.rb 0.9.0 - Toolkit for LLms with support for chat, streaming, tool calling, audio, images, files, and JSON Schema generation.
llm-shell 0.6.0 - Extensible, developer-oriented console for LLM communication.
lluminary 0.2.4 - Framework for building applications with AI that can define tasks, manage prompts, and handle LLM interactions.
n2b 2.0.0 - Transform development workflows with intelligent command translation, code analysis, and seamless Jira/GitHub integration.
nitpicker-code-review 1.3.0 - CLI that uses AI to review your staged Git changes and provide constructive feedback.
numo-liblinear 2.4.0 - Binding to the LIBLINEAR library or large-scale regularized linear classification and regression.
ollama-ruby 1.2.1 - Interact with the Ollama API.
omniai 2.8.2 - Standardizes the APIs for multiple AI providers.
omniai-tools 1.0.0 - Pre-built tools for integrating common tasks with OmniAI.
openai 0.7.0 - Official Ruby library to access the OpenAI API.
ox-ai-workers 1.1.3 - State machine with OpenAI generative intelligence integration.
prompt_manager 0.5.6 - Manage prompts for a generative AI system.
rails-mcp-server 1.2.1 - Implementation of Model Context Protocol server for Rails projects.
raix 1.0.1 - Add discrete LLM AI components to your Ruby applications.
roast-ai 0.3.1 - Convention-oriented framework for creating structured AI workflows.
rover-df 0.5.0 - Data frames for Ruby.
ruby_llm 1.3.0 - Interface to interact with modern AI models. Chat, generate images, create embeddings, and use tools.
tiny_mcp 0.2.0 - Implementation of the Model Context Protocol to create and serve tools locally for AI assistants.

New Gems

Links to RubyGems page.

rag_embeddings - Manage AI vector embeddings in C with Ruby integration.

internator - CLI that automates iterative pull request improvements using OpenAI's Codex. Cycle through objectives, make incremental changes, automatically commit and push each update.

agentic - a CLI tool to build and run AI Agents in a plan-and-execute fashion.

shared_tools - Collection of local tool definitions for AI libraries to use in generative AI applications.

contai - Automatically generate content for your Rails models.

sentinel_rb - LLM-driven prompt inspector designed to automatically detect common antipatterns in prompts before they reach production systems.

riktoken - Pure Ruby implementation of OpenAI's tiktoken library for BPE tokenization.

cv-parser - Extracts structured information from CVs and resumes in various formats using LLMs.