Genkit

The Genkit Dart community is building some incredible things 🙌 Jhin Lee just shipped genkit_llamadart — a Genkit plugin that brings local LLM execution to Dart using llama.cpp. That means you can now run Llama models entirely on-device, with full privacy, offline support, and zero API costs. Here's what makes this exciting: 🔒 Full privacy — your data never leaves the device 📴 Offline-first — works without an internet connection 💰 Zero API costs — no cloud, no billing surprises 🔌 Seamless integration — same Genkit API and plugin system you already use This is exactly the kind of community contribution that pushes Genkit Dart forward. Check it out 👇

Build your own AI search CLI with Genkit! 🔍 Xavier Portilla Edo has written a fantastic tutorial on building an interactive CLI chat tool powered by Genkit, Gemini, and Tavily. Here's what you'll learn: 🤖 How to define a web search tool with Tavily 💬 Building an interactive agentic chat loop with Genkit 🛠️ Using the Genkit Dev UI to debug your agents in real time It's a great end-to-end example of Genkit's tool-calling and agentic capabilities applied to a real developer task. Read the full tutorial: https://lnkd.in/gZTxKfAq

🚀 Genkit Dart v0.12.0 is here! We've shipped a new release for Dart developers with exciting improvements to our schema library and provider support. ✨ Schemantic now supports bring-your-own-type, giving you full flexibility over schema definitions 🛠️ Upgraded Anthropic & OpenAI SDKs for more robust provider support across platforms ⚠️ Breaking change: The middleware tool hook now returns Part for greater composability in complex pipelines Check the full changelog for details and migration guidance: https://lnkd.in/g9asQYfN

We love seeing the community's reaction to new Genkit features! 🎉 Developers across the ecosystem are already exploring and sharing their thoughts on Genkit's latest capabilities. It's incredibly motivating to see the community engage with what we're building. Check out what the community is saying: https://lnkd.in/gKguPHtB

Stop juggling multiple API keys. Use one Genkit plugin instead. 🔑 This blog post by Alexander Nohe is a great guide to using Vertex AI's Model Garden plugin in Genkit to access Claude, Mistral, Gemini, and Llama from a single plugin — no separate API keys, no separate consoles, no quota headaches. It's a practical, developer-friendly approach to model flexibility, and a fantastic example of the content our community creates. Read the full walkthrough: https://lnkd.in/gvgcESbb

The Genkit ecosystem continues to grow! 🚀 We are proud to see the community-built Genkit Java SDK featured on DZone. Xavier Portilla Edo has written a comprehensive guide on building generative AI applications using Java and Genkit. This is a testament to the power of the Genkit community in bringing modern AI tooling to every developer, regardless of their preferred language. Check out the article on DZone: https://lnkd.in/g3f4RzxZ

How do you align on prompting for production? 🤝 Effective prompt engineering isn't just about the magic words; it's about the workflow. Google Cloud Samples team at Google Cloud shared how they align on prompts using Genkit to ensure consistency and quality across their entire development team. It's a great look at how Genkit helps teams collaborate on AI-powered features. Read the full article on Dev.to: https://lnkd.in/gEjNxjkx

Building with Gemini at scale? 🚀 The team at Google Cloud just published a fantastic deep dive into how they use Gemini to generate code samples at scale. They share 7 technical takeaways, including how they leverage Genkit for prompt management and systematic evaluation to ensure quality and consistency. If you are moving AI features from prototype to production, this is essential reading. Check out the full post on the Google Cloud blog: https://lnkd.in/gdJQ_fTA

🚀 Genkit 🤝 AWS Lambda We are excited to highlight the new AWS Lambda integration for Genkit, built by community member Xavier Portilla Edo. Genkit is designed to be platform-agnostic, and this plugin makes it easier than ever to deploy your generative AI workflows to AWS. This is a great example of how the Genkit ecosystem is expanding beyond Google Cloud and Firebase to empower developers wherever they choose to build. Check out Xavi’s guide and the repository here: https://lnkd.in/g5rjbTET

Genkit Go v1.5.0 is out! 🎉 We've shipped a solid release for Go developers with new APIs, expanded model provider support, and important reliability fixes. What's new: ✨ Define flows that stream responses in real-time via Go channels. 🌊 Channel-based streaming API for building real-time AI pipelines 🧩 OpenAI structured output support in the compat-oai plugin 🤖 Anthropic native structured output support in the Anthropic plugin 🔧 Vertex AI and Anthropic model resolution improvements A huge thanks to community contributors apascal07, traut, hugoaguirre, dmeremyanin, and Zereker for their contributions to this release! Check out the full release notes: https://lnkd.in/gNyv38fP