TOVIO revisits the foundations of version control for teams where people and AI agents work together. These are the mechanisms behind that approach.
AI agents need their own authority.
TOVIO gives agents signed, expiring capability tokens with explicit path and operation scope. Agent-authorized work carries provenance. Available transcripts can add review context, with observed capture clearly labelled as unattested and self-reported.
A repository can contain different read boundaries.
Policy-protected file content is encrypted during snapshot construction. Copies of those protected objects remain ciphertext for readers without entitled keys. Unprotected content, metadata, and authorized working copies have different boundaries.
A stable Change ID lets supported rewrites preserve the identity of the work. Conflicts are stored as typed objects, keeping the disagreement available for inspection and resolution. Native binary storage brings code and large assets into the same history without LFS.
These are product capabilities under documented supported profiles, not claims of general availability, independent certification, or measured customer outcomes.
A description you can use.
For an article, show notes, a company listing, or an introduction. These descriptions are editorial background, not attributed founder quotations.
In one sentence
TOVIO is an open-core version control system for developers and AI agents, with cryptographic per-file permissions and stable change identity.
About TOVIO
TOVIO is an open-core version control system built for developers and AI agents working together. It combines cryptographic per-file permissions, scoped agent authorization, stable change identity, first-class conflict objects, and native binary storage. Its engine, CLI, SDKs, and specification are Apache-2.0; the Forge collaboration product is commercial. TOVIO was founded by Dustin Williams. Learn more at tovio.dev.
Founder background
Dustin Williams is the founder of TOVIO, an open-core version control system for developers and AI agents. His work on TOVIO focuses on explicit access boundaries, attributable agent work, and preserving the identity of a change as it evolves.