A local model can be useful when you want to try inference on your own machine, but installing an agent app does not automatically install a model. Start by checking the model service, your hardware and the other connections your task needs.
Find the model, then test it
The APX model guide describes automatic discovery of Ollama and LM Studio on the local computer. Their available chat models appear as local choices; if the app closes, a model can stay listed as unavailable. A local option still needs a suitable model running on the machine where it is discovered.
Pick one agent and a harmless sample task. Ask for a short answer with sources or a checkable calculation. Note the time, result quality and any error. Repeat the same task with another model using the same inputs, and inspect the outputs rather than judging by fluent wording alone.
Check what else leaves the machine
Local inference does not automatically make browser searches, integrations or fallback providers local. Review the selected model, fallback chain and connected tools before using confidential material. A fallback can have different data practices and costs. If a model is unavailable, check the model application and APX’s engine settings before assuming the task completed.
Sources
More articles
How to let an AI agent work on a site without sharing your password
A practical guide to browser handoffs, sessions and safer sign-in for AI agents.
Can an AI agent work from WhatsApp or Telegram? What to check first
Plan a chat-based agent workflow around identity, approval, delivery and privacy before connecting a messaging app.
Why an AI agent cannot use a connected tool: an MCP checklist
Troubleshoot an MCP connection in stages: scope, approval, sign-in, tool discovery and a safe test.