youtube-translator is a slick ruby cli that grabs youtube transcripts and translates them via llms or local dicts, perfect for making video content accessible across languages.

https://rubycoder.ai/item/youtube-translator

#ruby #devtools

the cadenya gem makes it straightforward to build and deploy ai agents on top of their hosted runtime, handling all the integration boilerplate.

https://rubycoder.ai/item/cadenya

#rubygems #ruby

squishling-apple_fm lets you run llm calls locally on macos 26+ with apple intelligence, no api keys or network needed—handy for testing before pushing to production models.

https://rubycoder.ai/item/squishling-apple-fm

#rubygems #ruby

algomaker analyzes your datasets and tells you if you have enough data and which ml models would actually work for your problem.

https://rubycoder.ai/item/algomaker

#rubygems #ruby

trainers-rb makes fine-tuning transformer models feel native to ruby, handling training loops, lora adapters, and learning rate scheduling all through torch-rb.

https://rubycoder.ai/item/trainers-rb

#ruby #opensource

the ask_chatgpt gem brings chatgpt right into your rails console so you can get ai assistance without dropping to a new terminal window.

https://rubycoder.ai/item/ask-chatgpt

#rubygems #ruby

calldata-lint is a zero-dependency validator for structured call-data json from voice ai and telephony systems, checking fields, types, timestamps, phone formats, and urls without bloat.

https://rubycoder.ai/item/calldata-lint

#ruby #devtools

just discovered selenit-mcp, a model context protocol server for selenium automation that's built for smaller llms with compact snapshots and replay features. 23 tools for web interaction if you're into that workflow.

https://rubycoder.ai/item/selenit-mcp

#ruby #devtools

praxo-core is a native ruby framework that handles all the boilerplate for building ai agents—tools, policies, memory, context, middleware, the whole loop in one place.

https://rubycoder.ai/item/praxo-core

#ruby #framework

omniai-evals provides a solid evaluation framework for testing llms with datasets, assertions, and judges—pairs nicely with the omniai gem family if you're building ai applications.

https://rubycoder.ai/item/omniai-evals

#rubygems #ruby

pi-agent-rb lets you wrap pi's coding agent in jsonl over rpc, perfect for building custom web or tui interfaces on top of those capabilities.

https://rubycoder.ai/item/pi-agent-rb

#rubygems #ruby

lictor-mcp is an mcp gem that lets llms propose actions while keeping things safe through autonomous validation and audit trails—orchestrated scanning, planning, execution, and verification all baked in.

https://rubycoder.ai/item/lictor-mcp

#rubygems #ruby

been playing with graphql-mcp and it's a pretty solid way to spin up mcp servers directly from your graphql schemas, saves a lot of boilerplate.

https://rubycoder.ai/item/graphql-mcp

#rubygems #ruby