Local models plug into Claude Desktop via Ollama
Claude Desktop can be pointed at Ollama as a third-party gateway, so a prompt can run on a model sitting on your own computer instead of a cloud endpoint. Ollama says the setup takes a toggle: open Ollama, select Claude, switch it on, and Ollama configures the gateway itself.
While connected, you can pick any model inside Ollama, local or on Ollama's cloud. Turning Claude off in Ollama restores the previous setup, and Anthropic models stay available through the same toggle.
Ollama says prompts and data don't go to Anthropic or Ollama, telemetry is disabled by default, and its Zero Data Retention policy covers all of its models and services, cloud or local. Local models get configured on the Settings page.
Related stories
- Anthropic's OSS Scanner skips human review on bug reports
- Claude Max subscribers get up to $200 a month for the API
- AWS's Strands Decider 2B picks from your list, never writes
- Claude's hillclimb cut support costs to about a fifth
- Claude Directory opens a submission flow for plugins
- Four Opus 5.5 API changes make Opus 5 requests fail
Comments
No comments yet. Be the first.
Join the conversation
Sign in with Google to leave a comment. Your name and avatar come from your Google profile, and the comment appears after moderation.