Frequently Asked Questions

HYVE Ether OS is a desktop operating system built around AI agents as first-class citizens, not bolted-on browser tabs. The substrate runs your AI workloads locally where possible, encrypts everything against quantum-capable adversaries, and gives you 48 technologies to compose with — Omega for orchestration, Tide for cognitive sensing, Augur for security, Studios for creative work, and more.

Anyone who wants AI to work for them without surrendering their data, identity, or leverage to a single cloud vendor. Day-one targets are developers (the MCP Bridge integrates with Claude Code, Cursor, Zed, Continue.dev), creators (Studios for image, audio, video, motion capture, sensing), and privacy-conscious power users who want every default flipped toward sovereignty. Civic and federal-compliance customers are a Wave 3 audience.

Those operating systems were designed around applications you launch and quit. HYVE Ether OS is designed around agents that watch, decide, and act over time. The substrate has native primitives for the things agents need — provenance-keyed sandboxing, deterministic decision orchestration, encrypted inter-agent messaging, ambient context awareness — that legacy operating systems either bolt on as afterthoughts or do not provide at all.

HYVE Ether OS uses a Linux kernel base for hardware compatibility, but the entire userspace shell, agent substrate, and integration layer is purpose-built. From a user's perspective it is its own OS — you do not interact with the underlying kernel except through HYVE-controlled surfaces.

The standalone SDKs (Tide, Augur, Narrator, Trust, Omega-core) are open-source-licensed for personal use. The OS shell, the substrate primitives, and the integration layer are commercial software — your closed-beta license entitles you to use them for personal or internal-business use under the terms at /legal/terms.

The interactive demo at /try runs an in-browser facsimile of the Omega Shell with five surfaces (Omega, Nexus, Tide, Sentinel, Settings). It is not the live OS — that shipped at beta on July 1, 2026 — but it accurately represents the surface, the navigation model, and the data shapes you will see on first boot. The full feature tour, with real screenshots from a running install, is at /os.

HYVE Ether OS is in closed beta and not for public sale right now. Closed-beta members accept the risk that beta software is incomplete and may evolve. Access is by request — request access at /beta.

No. HYVE Ether OS is not a subscription. Closed-beta access grants a license that is yours, including the updates we ship during the beta. Stop renting your intelligence.

HYVE Ether OS is in closed beta — there is no public sale right now. Access is by request at /beta. Closed-beta members get the downloadable OS and give feedback that shapes it.

During the closed beta there is no public sale. Where a purchase does apply, it is final and non-refundable except in narrow cases like duplicate charges, unauthorized payments, or license-not-delivered-after-14-days. The full policy is at /legal/refund.

There is no public sale during the closed beta. Access is a single-user license by request; multi-seat licensing for organizations will be available at public launch through Enterprise contact.

Yes — when you request access at /beta, enter the recipient's email address. The license key is bound to the email used in the request, not to the requester.

There is no public sale during the closed beta, so there is nothing to pay for right now. Request access at /beta.

There is no public sale during the closed beta. Sales-tax handling is being added before public launch.

No. The OS does not send telemetry, usage data, or activity logs. Operator state lives in ~/.hyve/ on your local machine, encrypted with vault-derived keys. Optional features (cross-device sync, marketplace listings, crash reports) explicitly opt-in to relay endpoints, and where they do, payloads are end-to-end encrypted before they reach our servers.

No. The website is a relay-and-attestation backend. Where OS-side features transmit your data — sync envelopes, inter-user messages, collab rooms — the data is encrypted client-side using ML-KEM-768 + ChaCha20-Poly1305 before transmission. We cannot decrypt the payloads even under subpoena.

The bus, the storage, and the relay all use ML-KEM-768 — a NIST-standardized post-quantum key encapsulation mechanism — from the first byte of any session. Combined with 512-byte uniform cells (HYVE-UT) and Ed25519 signatures, this means traffic captured today is not readable when cryptographically relevant quantum computers arrive. Most software defers this to 'when it matters'; we encrypt against that future starting today.

The camera, microphone, and other privacy-sensitive peripherals are gated at the kernel module level — below userspace, below the browser, below any application. When you flip the kill, no software can override it. This is enforced at the lowest layer of the operating system.

HYVE uses Shamir 3-of-5 social threshold recovery. You designate five trusted contacts during setup; any three of them can collectively help you reconstruct your vault. No single contact ever has full access to your data. We cannot recover your vault for you — that is the cost of true sovereignty.

A second PIN you can configure at setup. Entering this PIN unlocks the OS into a clean decoy state while irreversibly wiping the real vault — physical-coercion resistance built into the unlock flow.

SOC 2 Type II is in preparation, not yet certified. Several controls are implemented today (type-system-enforced wire shapes, per-tenant key isolation, parameterized SQL, OsRng). A full third-party audit is scheduled post-launch. The /security page details our current posture and the certifications we are working toward.

HYVE LLM Creator keeps your intelligence yours. You bring your own keys — HYVE never sees your password — and orchestrate through Omega, with your models running on your own hardware. Nothing leaves the machine unless you explicitly export it.

Yes. The local-engine category includes self-hosted LLM runtimes that consume your GPU directly. HYVE will function fully — Omega, Augur, Tide, Studios — without any cloud LLM provider configured. Frontier models are an option, not a requirement.

Omega is the decision orchestrator. It takes four signals — Tide (your cognitive state), Trust (the requester's reputation), Narrator (the action's audience tone), Augur (security analysis of the action) — and runs a deterministic decide() function that returns refuse / clarify / proceed. The function runs in under 200 nanoseconds per rule, which means every agent decision can be replayed exactly in Observatory for audit.

Only if you grant them that permission. Every agent action passes through Omega's decide() gate. You configure the policies; you see every decision in Observatory; you can refuse, clarify, or revoke any agent at any time. Agents have no privileged kernel access — they run in Leaper's differential sandbox tier appropriate to their provenance.

Augur's static analysis runs on every agent-generated artifact (code, configurations, scripts) before they execute. Trust tracks the historical accuracy of each agent. Narrator's render templates produce structured output, not free-form prose. Together these make AI errors visible and constrained — but no AI is perfect. HYVE's role is to make the failure modes detectable and reversible, not to claim infallibility.

Discrete AI units with a defined capability surface, a permission set, and a trust score. An agent could be 'summarize meetings,' 'monitor security feeds,' 'manage a creative project.' They live in your machine, communicate over the bus, and are composed via Omega. The Nexus marketplace will host third-party agents you can install.

Probably yes if it is a 2018-or-newer x86_64 laptop with 8 GB+ RAM and an SSD. The /hardware page lists tested machines and explicit minimums. If your machine is on the unsupported list (Apple Silicon, ARM mobile, pre-2015 CPUs), do not buy early access until the corresponding roadmap phase.

Intel-based MacBooks (pre-2020) — yes, with caveats around T2 chip and proprietary peripherals. Apple Silicon (M1/M2/M3/M4) — no, not at beta or public launch. ARM64 Apple Silicon support is a post-launch roadmap phase.

Beta members receive an ISO download link by email on July 1, 2026. Flash the ISO to a USB drive (Rufus on Windows, dd on Linux, balenaEtcher on any), boot from the USB, and follow the installer. Full step-by-step instructions ship with the email.

Yes. The installer offers a 'install alongside existing OS' option that resizes the existing partition non-destructively (with backups still recommended). UEFI or Legacy BIOS both work.

Yes for the shell — VMware, VirtualBox, QEMU/KVM, Proxmox, and Hyper-V all work. Attach the ISO as a virtual CD/DVD (no USB needed); either EFI/UEFI or BIOS firmware boots it, with Secure Boot off. Give it ~105 GB disk and 8 GB+ RAM. Studios and Sense will not function fully in a VM (they need real GPU and WiFi-CSI access). Apple-Silicon Macs cannot run it in Parallels/UTM (x86-64 only) — use an Intel/AMD host. For full-fidelity testing, a bare-metal install is recommended.

Yes — and the SDKs are available now. Tide, Augur, Narrator, Trust, and Omega-core all have Rust, TypeScript, and Python SDKs. Build agents on your existing development workflow, sign them with your operator key, and submit to the Nexus marketplace.

Developers list agents, organs, themes, recipes, extensions, or program plugins. Operators install them through the OS's Nexus surface. Pricing is set by the developer; HYVE takes a small platform fee. The first marketplace listings ship at beta.

Six: agent, organ, surface, studio extension, connector, theme. Each has a manifest.toml describing capabilities, permissions, and the install surface. Plugins run on ports 7900-7999, sandboxed by Leaper's differential policy.

Per-product SDK docs are at /tide, /augur, /narrator, /trust, /omega. The architecture and security model are documented at /security. Deep API references will be published before public launch.

Yes. The SDK wire formats and core APIs are versioned and will not break between beta and public launch. Where breaking changes are necessary, they will land behind feature flags with deprecation windows of at least 6 months.

Check spam first. Then use the order-lookup tool at /order-lookup with your email and Stripe/Coinbase session ID — it will resend the email. If still nothing after 14 days, email Dev@vibesoftwaresolutions.com with your session ID and we'll resolve it personally.

One inbox handles everything: Dev@vibesoftwaresolutions.com — general support, security reports, legal questions, or reaching Anthony directly. We aim for a 48-hour response.

No. The license key never expires. You can redeem at any point after beta went live on July 1, 2026. Save the key somewhere safe — your password manager is recommended.

Yes — your license is bound to your email and operator identity, not to a specific machine. You may install on multiple machines you personally own. Multi-user / multi-seat installation is an organization plan, available at public launch.

Closed-beta members get the source code archive of the OS shell (under a perpetual-use license) and the substrate libraries. You retain the right to run your installation indefinitely. This is a contractual commitment in /legal/terms.

The closed beta went live on July 1, 2026. Closed-beta members received a download link by email that morning.

We haven't set a public-launch date yet — we're focused on the closed beta and will announce it when we're confident. The full roadmap is at /roadmap.

During the closed beta — July 24, 2026 onward — the OS is downloadable to members and their feedback shapes it in real time. On a date to be announced we reach full public launch. The full roadmap is at /roadmap.

Mobile companion apps are a post-launch roadmap phase. The OS itself runs on x86_64 desktops and laptops, and already includes on-device cross-device pairing and peer-to-peer LAN file transfer — no central server — for the machines you own. The iOS/Android companion apps that extend this to phones, connecting to your HYVE installation for relevant features, are the post-launch phase. No date committed yet.

Apple Silicon (M1/M2/M3/M4) support is a post-launch roadmap phase. Not at beta, not at public launch. We will announce when work starts.

Three named research lines: hardware accelerator partnerships for the on-device LLM stack, deeper civic and federal compliance certifications (SOC 2 Type II audit, FedRAMP path), and the federated agent reputation network across HYVE installations. Early-access members will be polled before any major direction commits. /roadmap has the full picture.