The wait is finally over. The most highly requested feature has landed! 🐧

🌍 1. Hello, Linux!

FastFlowLM is officially available natively on Linux!

The best part? It uses the exact same recipe you already know and love on Windows.

No new workflows to learn—just seamless, cross-platform power ready to scale.

💻 2. Cross-Platform Hardware Validation

Take the guesswork out of your setup.
We've rolled out comprehensive hardware validation for both Windows and Linux environments. FastFlowLM now automatically ensures your system is fully optimized and ready to run before you even start.

flm validate

🛠️ 3. Improvements & Bug Fixes

We’ve packed this release with highly requested features and critical fixes to keep your workflows flowing smoothly:

✨ What's New

• JSON Output: Full support for structured JSON output to keep your downstream applications and agents parsing perfectly.
• Smarter Embeddings: Upgraded our embedding logic to process all array elements, rather than just the first one.
• Model List Refined: Improved handling of the model list.

🐛 Bug Fixes

• CPU Spikes: Squashed a bug causing unexpected 100% CPU usage in some downstream apps.
• Prompt Caching: Fixed a prompt cache issue to keep your generations lightning-fast and efficient.

🙌 4. Acknowledgements

A massive shoutout to @superm1 , @jeremyfowers, and @jgmelber! Thank you for your insights and contributions to making this monumental Linux release a reality.

As always, thank you for using FastFlowLM!
More speed. More models. More flow. 🚀