# DVIDIA

> Human skill. Open intelligence.
> Applied AI skills and open-weight inference under one workspace.

## Current status (2026-09-20)
- Early preview. Do not describe paid services, payouts, robot task success, a token, or a live order book as available.
- Browser recorder captures one camera without audio, with local video storage, playback and original download on supported browsers. Physical-device coverage remains a release gate.
- Drafts and recordings are local to the browser. Saving a draft is not cloud publication.
- Native dataset target is LeRobot v3. Existing experimental exports have not passed upstream loader validation. Raw video downloads are not LeRobot datasets or trained policies.
- Inference onboarding/routing/metering is in development. No live DVIDIA completion endpoint is advertised.
- Premium workspaces will sell defined services. Contributor shares must be backed by collected service revenue. Free downloads do not produce earnings; no premium income rights or guaranteed returns are sold.
- Main code and standalone CLI remain private previews. Two CLI implementations need consolidation before public installation instructions.

## Principles
- Open, reusable artifacts with explicit per-artifact licenses and provenance.
- Models retain their individual licenses. Open weights do not automatically mean unrestricted use.
- Competitive execution with measured price and service quality.
- Distinguish recording, dataset, policy, and hosted service.
- Use actual evidence; no invented usage, benchmark, payout, or provider numbers.

## Pages
- https://dvidia.org/ — product overview
- https://dvidia.org/store — illustrative skills and source dataset references
- https://dvidia.org/record — recorder
- https://dvidia.org/workspace — local drafts and planned workspace services
- https://dvidia.org/serve — inference roadmap
- https://dvidia.org/docs — implementation status
- https://dvidia.org/spec — target artifact contracts
- https://dvidia.org/manifesto — product thesis
- https://dvidia.org/ship — acceptance milestones
- https://dvidia.org/app — mobile status

## Brand
Cream paper, ink, voltage #3C6CFF, IBM Plex, visor mark.