This website
The pages of buildlys.com are static files. Hosting and a content delivery network process IP addresses and access logs for every visit, as any web server does. A visit counter from statinside.com records page visits.
When you press the download button, the site records the time, your IP address, your browser's user agent and the referral source, if there is one. It uses these records to count downloads, and then it redirects you to the installer. Nothing else is needed to download.
You need no account to read these pages or to download the setup program. Questions about the app go through the Send feedback button inside it, not through this website.
The VarintKeel app
VarintKeel runs locally on your Windows PC and sends no telemetry. Your chats, your working folders and your API keys stay on the PC. The app needs no account to run.
Secrets that appear in tool output are redacted before they reach the model. Prompts go only to the model endpoint that you pick. With a cloud model, your prompt and the file text the agent reads go to that provider; with a local model nothing leaves your PC.
Each action also appears in the chat as a card with its real result, so you can see which file was read and which command was run. An expanded card shows the full command output or the file diff.
Optional features that use the network
Web search is optional. When you use it, the query goes out to the web, and you can run it over an anonymizing network instead. The tool http_get fetches one page at a time when you ask the agent to read official documentation.
Control of chats from a phone browser is optional too. It needs one outbound connection from your PC, and it can be switched off.
Update checks tell the app when a new version is ready. The app then shows an Update button, and you decide when to apply it. A cloud model is the last network use: the endpoint you enter receives your prompts, as described above.
Your choices
You choose the model endpoint that processes your prompts and file text. A cloud AI provider, a model on your PC or the built-in model are all possible, and the Test button shows what an endpoint answers. This choice decides where your data travels.
You choose how much the agent may do. In Ask first it pauses before actions that change something, in Plan only it writes a plan and runs nothing, and in Auto-run it acts at once. Pick the mode that fits the folder you are working in, and use Auto-run only on copies.
You decide which optional features you use. Leave web search and phone control unused if you want the app to stay on your PC, and copy your server folder before any experiment, so that every change you try can be reversed.
Contact
Use the Send feedback button inside the app to report a problem, ask a question or suggest a change. The button is part of the main window of the app, so you do not need an email client or a separate form.
Describe what you asked the agent to do, what you expected and what the card showed. A concrete report is easier to repeat. Use the same button for privacy questions about the app or this website, and say which of the two your question is about.
Updated 8 November 2026
