OpenVision

English | 简体中文 | 日本語 | Français | Italiano | 한국어

See again what you once saw. See what others could see—with permission.

OpenVision is an open-source initiative for trusted first-person experiential media: capture, replay, private sharing, and provenance.

Why OpenVision?

Today's media can preserve images while losing the point of view, context, and trust information that make a personal experience meaningful. Recordings may also become locked inside one device, application, or account.

OpenVision explores a more human-centered path: help people preserve meaningful moments from their own point of view, revisit them later, share them intentionally, and understand where the media came from and how it changed.

The project does not claim to reproduce consciousness, read visual memory, or determine whether every piece of media is “real” or “fake.”

What we aim to build

• Capture: create first-person records with clear origin information.
• Replay: revisit an experience across supported devices and applications.
• Private sharing: choose who can access personal media and for how long, where technically possible.
• Provenance: preserve understandable information about origin and transformation.
• Portability: let people retain and move their records without permanent dependence on one provider.

OpenVision is in its founding stage. These are public goals, not claims of a stable specification or production-ready implementation.

Trust model

Verified provenance is not universal AI-fake detection. OpenVision uses three top-level states:

• Verified capture: participating systems can verify the declared capture origin and the history they attest to.
• Declared synthetic or enhanced: synthetic, generated, reconstructed, or materially enhanced content is clearly identified.
• Unknown provenance: origin or history cannot be sufficiently verified. Unknown does not mean fake.

Provenance supports specific, inspectable claims. It does not replace context, consent, journalism, or human judgment.

Principles

• Put human agency, dignity, and accessibility first.
• Make privacy and meaningful consent product foundations.
• Prefer private-by-default experiences and understandable controls.
• Separate source material from synthetic or enhanced derivatives.
• Keep records portable and trust claims honest about their limits.
• Build open foundations that can coexist with independent and proprietary products.

Directional roadmap

1. Foundation: establish shared vocabulary, ethics, governance, and interoperability goals.
2. Capture and replay: deliver a useful path for creating and revisiting first-person records.
3. Private sharing: improve selective access, portability, and transparent handling.
4. Ecosystem: enable interoperable implementations and, over time, broader open and decentralized participation.

Progress depends on demonstrated user value, safety, privacy, and trustworthiness—not only technical capability.

Open collaboration

OpenVision welcomes public proposals, research, documentation, reference work, and critical review. Shared interfaces and trust principles should allow open-source, commercial, and proprietary implementations to participate honestly in the same ecosystem.

Start with RFC 0000: Open Vision Manifesto. You can also read the project vision, roadmap, ethics principles, governance, and contribution guide.

License

Code and project documentation are released under the Apache License 2.0, unless a file states otherwise. Other assets may use separate, explicitly documented terms.