This website
This website is made of static pages. Pages are static. Hosting and a CDN process IP addresses and access logs so the pages can be delivered and kept running. A visit counter from statinside.com records page visits, which tells us how many people open the pages.
Pressing the download button records the time, your IP address, your browser user agent and the referral source, if there is one. These entries are used to count downloads. The button then redirects your browser to the installer. Counting is the purpose.
The website has no sign-in and no account. There is no account. You can read every page, press the download button and leave. Nothing on the pages asks for personal details, and no form collects anything beyond what is listed here.
The StatusRivet app
StatusRivet runs locally on your PC and has no telemetry. Nothing phones home. Your chats, your working folders and your API keys stay on the PC. No account is needed to run it.
Prompts go only to the model endpoint you pick. The provider sees the prompt. With a cloud AI provider, the prompt, including the text of any file the agent reads for that step, is sent to that provider and falls under its policy. With a model running on your PC, nothing leaves the machine.
Secrets that appear in tool output are redacted before they reach the model. Redaction happens first. The app also shows every action as a card with its real result, so you can see what was run and what was changed. Read the cards before you accept a diff, because the cards are the record of what the agent did on your machine.
Optional features that use the network
Web search is optional and can run over an anonymizing network. When you use web_search or http_get, the agent contacts the pages involved, so use them only for public documentation and not for secrets. Public pages only.
Control of your chats from a phone browser uses one outbound connection. It is optional. It can be switched off, and nothing in debugging depends on it. Leave it off if you work on a machine that must stay isolated, and use a model on your PC there as well.
Update checks use the network as well. The app shows an Update button when a new version is ready, and new versions are published automatically. A cloud AI provider also needs a network connection. The Update button appears.
Your choices
You choose the model endpoint. A cloud AI provider, a model on your PC or the built-in model are your options, and the prompts follow the choice you make. The models page explains each one, including what it means for your logs.
You choose how much the agent may do. Auto-run acts at once, Ask first pauses before actions that change something, and Plan only writes a plan and runs nothing. Pick the mode that fits the machine you work on.
You keep the work within your own systems. Scope is yours. Shell, web and SSH are not limited to one folder, so use StatusRivet only on software you wrote or are authorised to debug, on a lab or test machine.
Contact
Use the Send feedback button inside the app to report a problem or ask a question. It is the support channel for StatusRivet, and it reaches us from the app itself. Feedback is in-app.
The installer is not code-signed yet. Compare its SHA-256 checksum with the published manifest, and treat any Windows SmartScreen warning as a reason to check the file before you run it. Verify before you run.
This page describes the website and the app at version 2.9.26. Version 2.9.26 applies. When the way data is handled changes, this page changes with it.
Updated 2026-11-02
