Learn how the XA1 private infrastructure handles local execution, screen mirroring, and end-to-end user privacy.
Runs in browser. Holds chat cache, memory logs, and API keys exclusively in local secure storage. Never uploads them to XA1 servers.
`api.xa1.ai` processes requests statelessly. Chains inputs directly to GROQ/NVIDIA. Zero log retention. Serves as coordinates bridge.
`xa1_wireless_bridge.py` runs on your PC. Apple Shortcuts run on your iPhone. Actions execute inside your LAN boundary safely.
XA1 is a private, client-side personal AI agent platform. Unlike generic chatbots, XA1 integrates directly with your hardware. It runs on your web browser and connects to your local devices via a Python bridge helper or Apple Shortcuts to execute terminal scripts, manage local files, read clipboards, and visually coordinate clicks on your physical screens.
We follow a strict **Zero-Knowledge** model. Your API keys, settings, memory journals, and chat histories are stored solely inside your browser's local sandbox (`localStorage`). They are never uploaded, logged, or cached on our servers. When you issue a command, the cloud API statelessly compiles prompts to return cognitive coordinates, without retaining any trace of the data.
The companion script `xa1_wireless_bridge.py` runs locally on your PC. It establishes a sandboxed HTTP server on your local area network (LAN). High-risk operations (such as executing local bash scripts, reading filesystem logs, or controlling desktop applications) are sent from your browser to this local daemon. Because the connection is local, your system shell is never exposed to the public internet.
Apple iOS utilizes a strict application sandbox. To enable remote control of system features (Wi-Fi, Bluetooth, volume, screenshots, clipboard) without requiring jailbreaks, XA1 uses a native Apple Shortcut (`XA1 iPhone Master`). The Shortcut calls your unique API bridge endpoint statelessly, uploads current telemetry status, and dynamically executes native iOS system actions returned by the AI core.
Yes. XA1 does not manage or store your LLM provider API keys. Your keys (NVIDIA, GROQ, etc.) stay completely on your browser. When communicating with the models, the keys are passed directly in HTTP request headers from your client-side browser to the stateless model endpoint. No server intermediate is ever authorized to log or intercept your credentials.
No. XA1 preserves your local storage space. Chat history, visual screenshots captured from the iPhone screen mirror, and analyzed files remain strictly local. Visual analysis is processed instantly by the model and discarded. There are no backend database clusters storing user records or telemetry files.