Urgent.News

What's breaking now, across thousands of outlets.

Tech

How to Clean Up Ollama and LM Studio Models Taking Up Space on Windows in 2026

Your C: drive is red in File Explorer, and a good part of it is AI models you pulled into LM Studio and Ollama months ago. Each one is several gigabytes, they live in folders you never open, and neither app shows you everything in one list. Off Grid AI Desktop finds them for you, shows each one with its size, and lets you remove the ones you no longer use. LM Studio files go to the Recycle Bin,…

Your Windows PC's C: drive has a significant amount of AI model storage from LM Studio and Ollama. These models live in hidden folders you don't usually view, and neither app shows a consolidated list. To manage these files, use Off Grid AI Desktop, which identifies and displays each model's size. The app then allows you to delete unused models.

LM Studio places its files in %USERPROFILE%\.lmstudio\models, while Ollama stores its models in %USERPROFILE%\.ollama\models or a specified folder via the OLLAMA_MODELS environment variable. Off Grid AI Desktop reads these folders without altering the files, only moving or removing them as needed. The app also has a recycle bin for LM Studio models, and a safe way to remove Ollama models using the ollama rm command in PowerShell or Command Prompt.

Written by urgent.news from Dev.to's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.

Read the original at dev.to →

More in Tech

Black Friday Sale: System Design

Imagine, there are 4 database sharding strategies with 4 completely different outcomes. Pick the wrong one and your primary database melts on Black Friday.

  • Directory-based sharding recommended for Black Friday e-commerce
  • Hash and range sharding fail to handle high write/read volume
  • Caching directory in Redis mitigates potential bottleneck

Agents follow the runbook — including the step the attacker added

Agents follow the runbook — including the step the attacker added Give an agent a runbook and it will follow it. That is the point of a runbook.

  • Runbooks contain legitimate steps and a malicious step
  • TaintWall firewall blocks all 16 malicious cases
  • Firewall incurs 6-point utility cost

Fitting Three Objects in One Case: Modeling Multi-Item Packaging

The first article in this series modeled a product as a data model. The second one dealt with a curved object that broke the rectangular assumption.

  • Multi-item packaging case now uses layout with individual cavities for each object.
  • Cavity data structure includes id, inner dimensions, clearance, position, and orientation.

Implementing Consent and Account Deletion in a Fitness App

Implementing Consent and Account Deletion in a Fitness App A user disables analytics tracking in your fitness app. The settings screen shows that tracking is off, but your backend continues sending…

  • Define distinct processing purposes for user data, like workout tracking and analytics.
  • Create a ConsentStatus and ConsentPreference data model with audit history support.
  • Implement server-side preference updates with validation, persistence, and synchronization.

More from Thursday 29 October →