RoboRuby
Posts
Ruby AI News - June 24th, 2025

Ruby AI News - June 24th, 2025

RailsConf is coming to Philly soon, so you know, Go Birds!

Matt Solt
June 24, 2025

This edition features an entire Obie Fernandez AI newsletter, awesome new Active Agent docs, Rapid Ruby inspiring the next generation of Ruby AI developers, and much more.

Newsletter update: The full edition is now available inside your email client, so you can read through it without having to click through to the web. And if you are interested in Ruby AI and will be in Philadelphia for RailsConf on July 8th-10th and would like to meetup, please me know: [email protected].

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

Content

Announcement: Kody Kendall announced the release of llama_bot_rails, a Rails engine to turn your application in an AI agent powered by LangGraph and OpenAI. Explore your application code, query data, and run console commands with AI. There is also a short demo and a full walk-through.

Tutorial: Daniel Doubrovkine demonstrated Using Claude-Swarm to Upgrade Ruby Projects. Claude Swarm is a tool that orchestrates multiple Claude Code instances as a collaborative AI development team. The post details how Shopify is attempting to use claude-swarm to generate Ruby unit tests at some scale with an army of AI test agents.

Article: Dewayne VanHoozer published the Philosophy behind AIA, a Ruby gem command-line utility that facilitates interaction with AI models through dynamic prompt management. The article demonstrates how AIA takes a different approach to prompting by treating them as first-class executable scripts that can adapt, evolve, and experiment without touching your main codebase. The second tutorial in the AIA series details Mastering AIA’s Batch Mode: From Simple Questions to Complex Workflows. Future planned articles in the series include building AI workflows, interactive AI sessions, and extending AIA with custom tools and functions.

Update: Dewayne VanHoozer also updated AIA to version 0.9.7, bringing RubyLLM::Tool support, AI model management, enhanced executable prompts, and web content integration. Be sure to check out the AIA wiki and repository for more information.

Announcement: Scott Werner created Botrytis, an LLM-powered fuzzy matching tool for your Cucumber steps. The gem uses large language models to match semantically similar Cucumber steps, bringing flexibility to your tests even when they don't match exactly. Scott published a detailed write up of the new gem in The Moldy Cucumber Chronicles.

Tutorial: Exequiel Rozas demonstrated Adding llms.txt to a Rails application. The article explains the llms.txt specification, why should add it to your applications, and walks through a complete setup and integration into a sample Rails app.

Update: Carmine Paolino released version 1.3.1 of RubyLLM with fixes for OpenAI search parameters, Anthropic API errors, and add custom logger support. Carmine also highlighted Patrick Vice’s ruby_llm-mcp gem, which adds support for Model Context Protocol clients to use tools in RubyLLM.

Podcast: Valentino Stoll and Joe Leo hosted guest Scott Werner on the second episode of the Ruby AI Podcast 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 soon, will feature Obie Fernandez and look at Roast, the open-source framework for building structured AI workflows with Ruby.

Update: Sergio Bayona released VectorMCP 0.3.0, a server implementation for the Model Context Protocol (MCP) in Ruby. The release adds a two-layer validation system to secure MCP tools and prevent injection attacks, along with schema and input validation: X/Twitter post.

Tips: Kieran Klaassen continues to drop Ruby AI coding tips and tricks, including adding “bundle show” to your CLAUDE.md context file, using the stepper_motor gem to build AI workflows, and detailing his coding workflow on the podcast AI & I.

Article: Nityesh Agarwal published ChatGPT Taught Me Special Relativity - And Changed How I Learn, sharing the same personal insights and techniques he used to learn Ruby on Rails.

Launch: Alex Smith launched a beta of MVPrompt, an application to create a fully deployed Ruby on Rails app with generative AI. The project aims to allow both non-developer and developers users to prompt a fully-deployed Rails app with Claude and GitLab.

Tutorial: Otto Thornton-Silver looked at an AI enthusiast’s (failed) attempt at code porting using advanced agentic code gen completion with MCP tooling, rulesets, and meta-prompting in AI Agents Aren’t Ready For This (Yet).

Interview: Rhiannon Payne interviewed Irina Nazarova, organizer of the San Francisco Ruby Meetup and Conference, on Meet the Woman Leading the Ruby Revival in San Francisco. Rhiannon also sat down with Kinsey Durhan Grace, a senior engineer at Github, for a discussion on working with Copilot and AI.

Article: Rhiannon Payne also recapped The Best Way to Structure AI in Rails: MVC with Agents, Prompts, and Views (From SF Ruby). The recap looks at Active Agent, a Rails framework that offers a structured approach to building AI-powered applications through Agent Oriented Programming.

Launch: Kyrylo Silin launched WeUseRails, a showcase of web applications built with Rails, including an entire category for artificial intelligence applications. If you have an AI-powered Rails application, WeUseRails is free to join and submit your project. Kyrylo also started vibe coding a new game called Tech Debt. Can’t wait to see how it turns out!

Article: Giménez Silva Germán Alberto published PicoRuby: Bringing Ruby to Microcontrollers and the Edge of IoT. The post looks at PicoRuby, an mruby implementation designed to run on one-chip microcontrollers, useful for bringing Ruby AI applications to robotics and Internet of Things (IoT) devices.

Update: Vickash released Denko 0.15, an mruby electronics programming library to simplify working with embedded hardware. The source code is available on Github.

Tutorial: AI on Rails recorded a video on Building a Financial Portfolio Tracking Application with Cursor and Claude.

Update: Paweł Urbanek updated rails-pg-extras to add MCP integration, enabling PostgreSQL metadata and performance analysis with an LLM prompt. The gem and installation instructions are available on Github.

Tutorial: David Ang of Reinteractive created a guide for building a Virtual Assistant in Spree, aka Retrieval Augmented Generation. The example uses Ollama to run a model locally, Langchain.rb to interact with the model, and Qdrant as a vector database to store product embeddings.

Announcement: Martin Emde released studio and protocol-jsonrpc gems. Studio transforms any CLI command into a StdIO Model Context Protocol (MCP) server, exposing templated command-line tools. Protocol-jsonrpc implements the JSON-RPC 2.0 protocol layer to help with implementations of the Model Context Protocol in Ruby. The announcement and an example use case are detailed in this Bluesky thread.

Update: Kristoph committed an update to the intelligence gem to support OpenAI reasoning output by migrating the OpenAI adapter to the Responses API. This means the gem can now also support web search, image generation, computer use, and integration with the local shell. The announcement and a usage example are available in this X/Twitter thread.

Opinion: Rob Zolkos tested the Claude Code workflow previously shared by Kieran Klaassen in That Weird AI Workflow Might Just Work. In the piece, Rob discusses the importance of experimenting with new AI workflows, and dismisses the critics of Kieran’s workflow. Additionally, Rob posted How can we use AI without getting dumber?, positing that its up to developers to keep learning in the age of AI.

Tips: Levi Figueira published Rails AI-Assisted Development Guidelines, a set of best practices and development guidelines for building Rails 8 applications with AI agent collaboration using Claude.

Article: Ronak Bhatt looked at Exploring OpenAI’s o3-Pro Model: What’s New and Why It Matters, breaking down what the o3-Pro model is, what makes it different, and how to use it to power your workflows with the openai gem.

Tips: Daniel Tenner demonstrates his workflow and best practices for using Claude Code with Ruby on Rails in this X/Twitter thread.

Announcement: Brandon Shar created LLM-specific docs for InertiaJS Rails to use with Cursor or other AI-assisted coding agents: X/Twitter post.

Update: Charles Oliver Nutter pushed a fix for JRuby that enables all RubyLLM specs to pass for the Java implementation of Ruby.

Video: Thoughtbot live-streamed Can AI fix flaky Rails tests? In this session, Chad Pytel examines whether AI tools can help Rails teams detect, debug, and fix instability faster. Jimmy Thigpen of Thoughtbot also published Can AI rebuild a Rails page in Next.js? We tried it.

Documentation: Heroku added Ruby guides for managed inference and agent add-ons. The guides cover chat completions, embeddings, and image generations.

Tips: Nate Hopkins shared a cheat sheet covering Claude Code’s configuration files.

Article: Greg Molnar tricks LLM’s into doing something they shouldn’t be able to do with Exploiting LLM Chatbots.

Opinion: Justin Searls wrote Developers Can Either Be Scared of AI or Supercharged By It, arguing that even with AI’s current capabilities, its a huge boost for developers.

Article: Akshay Khot dispensed common patterns learned developing a Ruby on Rails AI training program in Working Effectively with AI as a Developer. The patterns include ensuring the quality of your prompts, having the model understand related code, and building incrementally.

Tips: Lucian Ghinda tweeted an example of prompting Claude Code to identify test cases for review before asking the AI to start writing test cases.

Article: Tommy Ball of Ruby on Rails development firm Viget published My Aha! Moment with AI with musings on AI tools, AI-first product design, and the happy moments of tool-usage-clarity in a digitally cluttered world.

Opinion: Breno Vasconcelos shared his thoughts on using RubyLLM to make integrating cutting-edge artificial intelligence into web applications simpler and more accessible with this LinkedIn post.

Article: JetRuby studied AI Agriculture - Future of Farming with examples of how AI is leveraging data to drive decisive actions across the agriculture technology landscape. Additionally, JetRuby published an AI Roadmap for AI App Development 2025.

Opinion: Valentino Stoll believes open source projects will soon have their own AI budget. Do you agree?

Wow, just think, some day soon every Github repo will just have an AI budget, and folks can literally watch their sponsorships add value through automated contributions to the maintenance and development of new features. Imagine being able to open an Issue and assigning funds to it!

https://x.com/thecodenamev/status/1935005914093769091

Events

Upcoming

June 24th - Meetup: 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.

June 28th - Conference: RubyKaigi Kansai in Kyoto, Japan will feature a talk by Katsuhiko Kageyama on World Side of Robot Connected by mruby and micro-ROS (translated) on June 28th.

July 9th - Conference: Day 2 of RailsConf on July 9th, in Philadelphia, Pennsylvania will include a “hack space” on Active Agent with Justin Bowen and Rhiannon Payne, designed to bring Rails developers and open source maintainers together for a day of hands-on collaboration.

July 16th - Meetup: 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 18th - Conference: RubyConf Africa includes a talk by Purity Maina on Your First Robotic Process Automation Bot in Nairobi, Kenya on July 18th and 19th.

July 19th - Hackathon: 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.

September 18th - Conference: 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

Scott Werner of Sublayer has made some incredible contributions to the Ruby AI ecosystem, including running ArtificialRuby and open-sourcing much of Sublayer’s source code. I wanted to resurface some Sublayer code that was previously highlighted by the Google AI for Developers showcase Enhancing AI Powered Developer Tools with Gemini API. In this example, any time a pull request is submitted to the Sublayer repository, an agent is triggered that updates the documentation based on the changes. The source code is available on Github, and there is also a corresponding Github Action you can use to trigger the agent.

New Versions

Links to release notes or changelog if available.

actionmcp 0.55.2 - Rails Engine with MCP compliant spec.
activeagent 0.4.0 - Agent-oriented AI framework designed for Rails, where agents are controllers.
aia 0.9.7 - CLI that facilitates interaction with AI models through dynamic prompt management.
anthropic 1.1.1 - Official Anthropic Ruby library.
askcii 0.2.0 - Terminal-friendly interface for interacting with LLM models.
baml 0.90.1-x86_64-linux - Interact with BoundaryML's language model clients.
boxcars 0.8.4 - Create new systems with AI composability.
decidim-ai 0.30.1 - Provides Artificial Intelligence tools for Decidim.
faiss 0.4.1 - Efficient similarity search and clustering for Ruby.
foobara 0.0.132 - Command-centric and discoverable framework with a focus on domain concepts and abstracting away integration code.
gitingest 0.7.1 - CLI that fetches files from GitHub repositories and generates consolidated text prompts for AI analysis.
intelligence 1.0.0 - Uniform interface for interacting with large language and vision model APIs across multiple vendors.
kerplunk-ai-prompts 0.1.33 - Collection of AI prompt templates for use in Kerplunk.
launchdarkly-server-sdk-ai 0.1.0 - LaunchDarkly SDK AI Configs integration for the Ruby server side SDK.
lightgbm 0.4.2 - High performance gradient boosting for Ruby.
memori-client 0.1.9 - Memori Client to interact with Memori backend and engine API.
n2b 2.1.0 - Development workflow with intelligent command translation, code analysis, and Jira/GitHub integration.
omniai 2.8.3 - Standardizes the APIs for multiple AI providers.
open_ai_bot 0.3.6 - Telegram bot for using ChatGPT, DALL-E and Whisper.
openai 0.10.0 - Official OpenAI API Ruby SDK.
or-tools 0.16.0 - Operations research tools for Ruby.
polars-df 0.20.0 - DataFrames for Ruby.
prreview 0.3.0 - Gathers details about a pull request for analysis by an LLM.
rag_embeddings 0.2.2 - Manage AI vector embeddings in C with Ruby integration.
roast-ai 0.4.2 - Framework for creating structured AI workflows.
ruby-kuzu 0.1.0 - Ruby bindings for the Kùzu embedded graph database.
ruby-mcp-client 0.7.1 - Integrate MCP servers to access and invoke tools from AI assistants.
ruby_conversations 1.1.10 - Library for managing AI conversations via a remote API.
ruby_llm 1.3.1 - Interface to interact with modern AI models. Chat, generate images, create embeddings, and use tools.
ruby_llm-mcp 0.3.0 - MCP client that seamlessly integrates with RubyLLM.
safetensors 0.2.0 - Simple, safe way to store and distribute tensors.
shared_tools 0.1.3 - Reusable tools using various LLM-provider API gems.
stigg-api-client 2.442.0 - Stigg API Client for Ruby.
sublayer 0.2.9 - DSL and framework for building AI powered applications through the use of Generators, Actions, Tasks, and Agents.
vector_mcp 0.3.0 - Server-side tools for implementing the Model Context Protocol.

New Gems

Links to RubyGems page.

desiru - Brings DSPy's declarative programming paradigm for language models to Ruby, enabling reliable, maintainable, and portable AI programming.

llama_bot_rails - Turn your existing Rails App into an AI Agent by connecting it to an open source LangGraph agent.

studio - Transforms any CLI command into a StdIO Model Context Protocol (MCP) server.

botrytis - Semantic matching for Cucumber step definitions using LLMs.

universal_document_processor - Handles document processing, text extraction, and AI-powered analysis for PDF, Word, Excel, PowerPoint, images, archives, and more with a unified API.

ruby_llm-schema - Ruby DSL for creating JSON schemas, used for defining structured data schemas for LLM function calling or structured outputs.

rails-pg-extras-mcp - MCP interface for rails-pg-extras, enabling LLMs to analyze the PostgreSQL metadata and performance metrics.

open_router_usage_tracker - Rails engine to log OpenRouter API usage and track user consumption over time.

llm_chain - Interact with LLMs through a unified interface, with native Ollama and local model support.

genie_cli - CLI that provides an interactive AI assistant to write tests, implement code, and manage your codebase with RubyLLM.

meta_workflows - Rails engine for creating and executing AI-powered workflows with human interaction points.

litellm_client - Client for the LiteLLM API.

asimov.rb - Ruby SDK for ASIMOV, a polyglot development platform for trustworthy, neurosymbolic AI.

react_to_rails - Converts React components to Rails’ ViewComponents using AI.

ai_pr_reviewer - Automatically review pull requests using AI to provide feedback, inline comments, and suggestions for improvement.

inplay_ai - InPlay AI SDK for Ruby.

micro_mcp - Framework for building MCP servers with a Rust extension that handles the lower level protocols.

miharu-ai - LLM token cost monitor.

things-mcp - MCP server for Things 3.

mcp-datetime-ruby - MCP server that provides datetime tools for AI assistants.

Jobs & Opportunities

Are you an organization searching for an expert Ruby AI developer, or a Rubyist looking for your next development role with AI? Please reach out and let me know the type of opportunity you’re pursuing: [email protected]