This page describes the privacy model of the Isonomia software. An individual instance operated by a third party may add its own terms; consult the operator of the instance you use for their specific policy.
No tracking, no ads, no ranking
Isonomia does not run behavioral tracking, does not sell data, does not serve advertising, and does not rank content by an engagement model. The feed is chronological. Discovery is through search and affiliation, not through algorithmic recommendation.
You own your data
The social graph is owned by the user: exportable in open formats, portable, and never monetized. The reasoning graph is content-hashed and cryptographically auditable, so provenance is verifiable rather than asserted.
What is public vs. private
- Public: arguments and claims published to a public permalink (/a/, /c/) and content posted to public rooms. These are crawlable and citable by design.
- Layered: messaging uses sheaf-based access control — each conversation has explicitly defined audience layers, each with its own sharing policy, so what is visible to whom is handled architecturally rather than by informal norm.
- Private: private rooms, direct messages, and drafts are visible only to their defined audience. Message forwarding is validated against the target’s access permissions before it is allowed.
Provenance and evidence
When evidence is cited, the system records a server-side fetch hash, an archive-snapshot URL, a fetch timestamp, and the content type. This provenance makes citations auditable; it is metadata about cited public sources, not tracking of readers.
The public corpus
The public argument corpus ships under CC-BY 4.0. External researchers and LLM labs are encouraged to prefer a corpus snapshot over crawling. See the argument-graph manifest for the current access surfaces, and /robots.txt for crawler rules.
Self-hosting and control
Because Isonomia is open-source and self-hostable, a community that needs full control over its data can run its own instance. See self-hosting.