Spatius turns real-time speech into a client-rendered avatar experience. Hour One focuses on AI video production, including virtual presenters, studio workflows, templates, and an Enterprise REST API that can automate videos by changing fields such as character, voice, transcript, titles, media, branding, and music.
A video can be unique to one recipient and still be generated before viewing. An interactive avatar must respond to events that were not fully known in advance.
| Decision area | Spatius | Hour One |
|---|---|---|
| Core workflow | Continuous real-time application session | Studio and API-based video production |
| Personalization source | Live user input, agent state, tools, and application events | Template variables and request parameters supplied before rendering |
| Visual output | Motion data rendered by AvatarKit on the client | Generated video with presenter, voice, media, transcript, and branding |
| API boundary | Speech-to-motion avatar layer around your AI stack | Official help describes an Enterprise REST API for automated video creation |
| Success metric | Responsive turns, interruption, recovery, and device behavior | Time to approved asset, template reuse, personalization throughput, and consistency |
| Best fit | Assistants, tutors, interviewers, kiosks, and interactive agents | Sales, training, communications, onboarding, and personalized video campaigns |
A personalized video can mention a customer's name, company, role, product, or account status without being conversational. It is rendered from known fields and distributed through a campaign, learning system, or application. A live avatar makes choices after the session begins and may call tools, ask follow-up questions, or change direction. Clarifying this distinction prevents teams from demanding real-time infrastructure for a batch campaign—or buying a render workflow for an interactive product.
Write one sequence diagram for each candidate using a real workload and real system owners.
The application captures a user turn, runs ASR, reasoning, tools, and TTS, then sends speech audio through Spatius to drive the avatar. AvatarKit renders the result while the session remains open for interruption and subsequent turns. Every component contributes to response time and recovery.
Hour One's official API description starts from a Studio template. A request can vary the character, voice, colors, titles, media, transcript, music, and logo. The service renders a video for review or distribution. The workflow should include data preparation, job status, retries, quality checks, storage, and delivery.
At enterprise scale, data quality becomes a creative-system problem. A misspelled name, broken media URL, unsupported pronunciation, or outdated offer can generate thousands of flawed videos. Build a preflight validator and a sample approval step before bulk production. For live sessions, equivalent safeguards belong in prompt evaluation, tool permissions, speech fallback, rate limiting, and escalation rather than a render queue.
Seconds, minutes, and hours describe different products even when the avatar looks similar.
Spatius is not designed to replace a template-driven video studio. Hour One's documented video automation is not equivalent to a conversational runtime that responds to unpredictable user turns. A hybrid workflow can use Hour One before a meeting or training module and Spatius during practice or support, but only if the handoff creates measurable user value.
Use one month of expected demand. Separate unique videos, unique live sessions, and views of already generated assets.
Last reviewed Aug 3, 2026. Confirm Enterprise API access, supported parameters, plan terms, and rendering workflow.
