AI integration

Published data, deterministic local graph tools, and an accessible browser workbench. No account, inference service, or backend API is required.

Uploadable AI bundle

One archive with the SQLite database, JSON and RDF data, zero-dependency Python and ESM libraries, an Agent Skill, examples, citation metadata, and both licenses.

Download AI bundle

Local workbench

Search, inspect direct relations, find authored paths, compute distinct closures, build subgraphs, compare concepts, and copy visible JSON—all in the browser.

Open workbench

Portable SDKs

Use the Python standard library plus SQLite, or a zero-dependency ESM module. Both preserve authored direction, source records, and content version.

Python SDK · ESM SDK

Agent Skill

Installable instructions for compatible agents. The bundle includes the skill next to its scripts, references, and local data.

Read SKILL.md

Static integrations

Small predictable JSON records, dereferenceable relation vocabulary terms, and a GET-only OpenAPI description support agents that can retrieve files but cannot run arbitrary code.

OpenAPI description · Relation vocabulary · Data manifest

Citation chain

mAtlas graph assertions need both their canonical concept or relation URL and their attached scholarly or reference sources.

Citation guide · CITATION.cff