Working on tool use and a better search!

LLM Controller CE v1.5 drops tomorrow morning! 🎉 This is a big one! Featuring an all-new Memory system, major UX improvements and polish throughout, and Benchmarking V2. One of our biggest updates yet!

Hard at work testing v1.5! We brought memory into v1.5! LLM Controller CE will remember you now!

If you have an @NVIDIA or @AMD GPU, or no GPU at all, you can run LLM Controller CE! Don’t want to set it up yourself? Point your AI assistant at our GitHub and let it handle the setup.

LLM Controller CE v1.4 is here. Local voice dictation. A better experience throughout.

🚀 LLM Controller CE v1.3.0 is live! Projects + Project Instructions, System Instructions, Docling document understanding, expanded image/file support, API Analytics, improved imports, security fixes, and a whole lot of polish. One of the biggest CE updates yet. 🔥

LLM Controller CE v1.3 is almost ready. 👀 Projects + Project Instructions. Personal System Instructions. Document understanding with Docling. API analytics. A much bigger attachment pipeline. Final testing is nearly done. One last audit, then release.

LLM Controller CE v1.2 releases tomorrow. 🚀 • Multimodal image support • Better attachment handling • OpenAI-compatible API • Better title generation control • Expanded Model Management #LocalAI #OpenSource #LLM

LLM Controller CE v1.1 is available now! github.com/tensioncore/llm-c…

Update for LLM Controller CE v1.1 will be on github tomorrow! github.com/tensioncore/llm-c…

LLM Controller CE v1.1 is in final testing. Plan to update main in a day or 2! URL IDs, title search, better UX in several drawers, auto start/stop logs+monitor, bug fixes. #opensource #localLLM

Don't I know it 😉

Yay for more excellent open source!

We run our production server with 2x V100s and get 160TPS+ on GPT-OSS 20b

Have you tried LLM Controller? We run 2x V100s and get 160TPS+ with GPT-OSS. Our server costs about $25 per month in power.

We're working on solving whatever problems are left 😎

I know an excellent way you can run 600B models locally 👀

UI improvements, Search and RAM usage. LLM Controller CE v1.1 is in production testing.