Connect an AI Model to SentryEdit

SentryEdit does not lock you to one model. You pick a cloud AI provider, a model running locally on your PC, or the built-in model, and the agent uses it to plan edits to your Verse files. The choice decides where your prompts go.

Three ways to connect

A cloud AI provider

Connect a hosted model through a standard API endpoint. It suits large refactors that touch several .verse files, because big models handle long context. Your prompts travel to the provider you chose, while chats, project folders and API keys stay on your PC. The provider may charge separately.

A model on your PC

Run a model locally and point SentryEdit at its endpoint. It suits private projects and offline sessions, and nothing leaves your machine. Speed and quality depend on your hardware, so small edits and explanations of a single file are the best fit for a start.

The built-in model

SentryEdit includes a model that needs no API endpoint, so you can try the agent on a small Verse file right away. It suits first experiments with 1 file and 1 task. Check its answers like any other, and move to a larger model for bigger jobs.

Check the connection with Test

The Test button sends a short request to your endpoint and reports two numbers: the response time and the context window. The context window is how much text the model can read at once. Large .verse files and long chats need a bigger one, so check it before you start.

Choosing a model for the task

Use a bigger model for refactors, tricky logic and changes across many files. Use a smaller one for small edits, such as renaming a field or adding a Print line. SentryEdit can also send one hard step to a bigger model with run_on_model while the rest stays on your chosen model.

Quality varies from model to model. Read every diff, keep Ask first on, and test the result in UEFN. No model is right 100 percent of the time.