Carmine Paolino (@paolino) / X

Carmine Paolino

5,014 posts

Carmine Paolino

@paolino

Author of @chat_with_work and rubyllm.com | Previously @freshflowai, @turinginst & others | Produces music as Crimson Lake

Berlin, Germany

Joined October 2009

Pinned
Carmine Paolino
@paolino
Jun 9
RubyLLM 1.16 is out. - Concurrent Tool Execution - Rails-style Instrumentation - Custom Base URLs for Every Native Provider - Configurable Faraday Adapter - Deprecation Controls - Transcription Words and a deluge of fixes. github.com/crmne/ruby_llm…
00:00
5K
Carmine Paolino
@paolino
Apr 12, 2025
Happy to see RubyLLM hit 2K stars in just a month! A testament to how the Ruby community rallies around well-designed tools that make complex things simple. rubyllm.com
8.6K
Carmine Paolino
@paolino
Mar 25, 2025
Just ripped out Puma/SolidQueue/SolidCable for Falcon/AsyncCable/AsyncJobAdapter and holy smokes 🔥 The server's practically teleporting LLM responses now. No need for benchmarks to feel the difference. Massive props to @ioquatix for this async magic!
28K
Carmine Paolino
@paolino
Aug 21, 2025
RubyLLM is once again the number #1 trending repository for Ruby on GitHub!
3.3K
Carmine Paolino
@paolino
Jun 3, 2025
🎉 RubyLLM 1.3.0 is here! Just when you thought the DX couldn't get any better 😉 `chat.ask "What's this?", with: "diagram.png"` That's it. No more { image: "file.png" } Tons of other improvements too, like Ollama support, local config contexts, etc.
Release 1.3.0 · crmne/ruby_llm
From github.com
6.4K
Carmine Paolino
@paolino
Jul 23, 2025
Structured output just landed in RubyLLM! 🎉 No more parsing JSON and crossing your fingers. Define your schema, get exactly what you asked for. Every time. chat.with_schema(PersonSchema).ask("Generate a person") # => {"name" => "Alice", "age" => 30} rubyllm.com/guides/chat#st…
4.9K
Carmine Paolino
@paolino
Jul 9, 2025
Thread-based job queues are unfit for Ruby AI apps. Switch to async-job. 30 minutes. Same code. Same hardware. Tons more concurrent LLM chats. No max_threads. Oh, and async Ruby doesn't need any async/await library rewrites: RubyLLM gets faster for free
Async Ruby is the Future of AI Apps (And It’s Already Here)
From paolino.me
20K
Carmine Paolino
@paolino
Jul 16, 2025
Hey RubyLLM devs, I just released the Async guide on the RubyLLM website. Go make your apps faster and more scalable! :) rubyllm.com/guides/async
6.3K
Carmine Paolino
@paolino
Aug 23, 2025
RubyLLM is #1 on GitHub again! 🤩 Great to see lots of people showing confidence in Ruby + AI!
3.6K
Carmine Paolino
@paolino
Jul 21, 2025
RubyLLM is now 1 million downloads strong. Proof that the Ruby AI community is alive and well and keeps shipping! 💪
RubyLLM
From rubyllm.com
3.2K
Carmine Paolino
@paolino
Mar 19, 2025
"Ruby can't do concurrency well"? 🤨 Tell that to RubyLLM running parallel AI calls with proper streaming using Async: Async { 3.times { Async { RubyLLM.chat.ask("...") {|c| puts c} } } } w/o changes to the library. Guide soon at rubyllm.com
16K
Carmine Paolino
@paolino
Jul 23, 2025
Rails generators just landed in RubyLLM! 🚂 One command to set up everything: rails generate ruby_llm:install Creates your models, migrations, and initializer. Ready to chat in seconds. Thanks to @kieranklaassen for starting the PR! Get started: rubyllm.com/guides/rails#s…
3.6K
Carmine Paolino
@paolino
Mar 22, 2025
📣 Just added VCR support to RubyLLM! Now you can contribute without needing API keys for all providers. But don't worry - before each gem release, we still test against real APIs to ensure everything works. Lower barrier to entry, same quality assurance.
GitHub - crmne/ruby_llm: One delightful Ruby framework for every major AI provider. Build AI...
From github.com
3.8K
Carmine Paolino
@paolino
Mar 23, 2025
Just shipped RubyLLM 1.0.1! We've fixed temperature handling for OpenAI's o1/o3 models, improved tool naming, and added Rails system prompt examples. Also added base64 dependency for Ruby 3.4 compatibility. Making AI in Ruby elegant.
Release 1.0.1 · crmne/ruby_llm
From github.com
7.3K