Frequently Asked Questions

Yes. i4 Nexus never sees your conversations, your voice, or your memory. Your API keys and local data live in your OS-level vault (Windows DPAPI / macOS Keychain). When you use AI features, your requests go directly to the LLM providers you connect — through your own API key, under their privacy policies. Want total privacy? Connect your own or a local LLM, and nothing leaves your machine.

Everything stays on your machine. Brain² (the agent's persistent memory) is stored locally. API keys are stored in your OS credential vault. There is no external database holding your data — your device is the single source of truth.

No. API keys are stored in your OS-level secure vault and are only sent directly to the corresponding provider (Gemini, Claude, OpenAI, etc.) when a request is made. They are never sent to our servers.

No. The wake-word detection runs fully on-device and listens only for the trigger phrase — no audio is transmitted before the wake word fires. Once you speak a command, the audio goes to your connected LLM or voice provider through your own API key. The wake detection itself is entirely local; what happens after you speak is in the hands of the providers you choose.

Only you. Conversations are stored locally in Brain². We have no server-side access to your history. The only time any content reaches the internet is when you make an API call to a provider — and that goes directly to the provider, not through us.

Yes. You can wipe Brain² memory, revoke API keys, and uninstall the application at any time. Because everything is local, there is nothing on our side to delete — once you remove the application and its data, it is gone.

The standard plan is $19.99/month. The Early Bird plan — available only to the first 2,000 subscribers — is $7.99/month, locked for life. Beyond the subscription, you pay your LLM providers directly at their published rates through your own API keys.

The Early Bird plan locks in $7.99/month for life — that price never changes for active subscribers, even when the standard price rises to $19.99. It is available only to the first 2,000 subscribers and closes automatically when the limit is reached. Check the Pricing section to see how many spots remain.

BYOK stands for Bring Your Own Keys. You connect your own API keys from Gemini, Claude, OpenAI, or other supported providers. You pay those providers directly at their published rates. We never mark up, proxy, or see your API calls.

Most users spend under $5/month on API calls. The agent is designed to be efficient with token usage. You can also set a hard monthly ceiling — the agent stops making paid API calls when the limit is reached, preventing surprise bills.

Yes. Every paid plan starts with a 14-day free trial. You can explore the full feature set before any charge is applied. No credit card is required to begin the trial.

Yes. There are no long-term contracts. You can cancel at any time from your account settings. If you are on the Early Bird plan and cancel, the discounted price is lost — if you resubscribe, you pay the standard rate.

The agent pauses all LLM-dependent features gracefully and notifies you. Core offline features (wake word, local memory, device control that doesn't require an LLM) continue to function. You can reset your budget or raise the ceiling at any time.

Download the application for your platform, run it, and follow the setup wizard. You will be asked to add at least one API key (Gemini, Claude, or OpenAI). Once configured, say your agent's wake word and it is ready. Full setup takes under five minutes.

The core agent, wake-word detection, and Brain² memory all work fully offline. Features that require LLM inference (answering questions, writing, summarizing) need an internet connection to reach your API provider. You can use the agent in a hybrid mode — offline for device control, online for intelligence.

Brain² is the agent's persistent local memory system. It stores notes, preferences, routines, and context across sessions. Unlike chat history, Brain² is structured and searchable — the agent actively uses it to personalize responses without you needing to re-explain yourself every time.

Yes. You choose the wake word, the agent's name, its voice (from available TTS options), and its personality profile. This is a genuine white-label experience — the agent can feel entirely custom to you.

Out of the box: answer questions, set timers, check the time and weather, control dashboard widgets, switch themes, play music via Spotify, search and control YouTube, open Google Maps, read news headlines, manage a to-do list, open applications, and shut down the system. More capabilities are added with each update.

The wake-word engine runs entirely on-device using a small local model. It continuously listens for the configured trigger phrase (default: the agent's name). No audio is streamed anywhere before the wake word is detected — it is similar to how Alexa and Siri handle this, but without cloud dependency.

Windows is available at launch with the full feature set. Linux is in active beta (Debian/Ubuntu/Arch). macOS is on the public roadmap. All paid subscribers receive each new platform update at no additional cost when it ships.

Windows 10 or later (64-bit), 8 GB RAM, and a microphone for voice features. A GPU is not required. An internet connection is needed for LLM-dependent features; core functions work offline. Exact requirements are listed in the download page.

Any model you have an API key for: Gemini (1.5 Flash, 1.5 Pro), Claude (Haiku, Sonnet), OpenAI (GPT-4o, GPT-4o mini), Mistral, and others. You can route different tasks to different models. The model selector is accessible per-agent in settings.

Yes. The agent supports smart home integrations via the Automation arm. Compatible ecosystems include Home Assistant, MQTT-based devices, and supported smart home APIs. The exact list of integrations is available in the documentation.

The current release is designed for single-user deployment with per-profile voice biometrics. Multi-user deployment is part of the Enterprise plan roadmap. If this is a requirement, contact us at support@i4nexus.com.

All active subscribers receive the new platform builds automatically at no extra charge. You will be notified by email when the beta opens for your platform. Linux subscribers can already join the beta from the account dashboard.