See app, model, hardware, and privacy changes that affect your setup.
Subscribe to local AI news with RSS
Latest update
Ollama's new decision endpoint turns short text into named choices and probabilities. Start with a local ticket-sorting request and inspect the answer.
·5 min read
Local AI Models
Read articleLocal AI Models
Splash brings model-specific Qwen inference to Bionic on compatible Macs. Check the hardware, install the engine, and select the matching local package.
·6 min read
Local AI Models
Liquid AI's smallest recent models target narrow jobs that fit ordinary hardware: extraction, tool routing, multilingual retrieval, and compact encoders.
·5 min read
Local AI Apps
Open WebUI 0.11 expands collaboration and background automation, while its security advisory gives production administrators a reason to update promptly.
·7 min read
Local AI Apps
Bionic is a separate agent application that can run locally or use linked and hosted compute; the original LM Studio app remains the low-level local model workspace.
·6 min read
Local AI Apps
Ollama still runs local models, but version 0.32 adds an agent whose default path can use hosted models and account-backed web tools.
·7 min read
Privacy & Security
A local model is not enough: local AI privacy depends on whether prompts, documents, embeddings, logs, tools, providers, and server settings also stay local.
·28 min read
Model Guides
A practical, hardware-focused guide to choosing the right Gemma 4 model for local chat, coding, document work, multimodal use, and experimentation.
·9 min read
Model Comparisons
Compare the exact download, available memory, text or image task, and app support before choosing among these three local models.
·8 min read