Skip to content

Docs · The platform

What Artifactum is

Artifactum is a governed memory for a company's facts, and the shared record its AI agents work through. The marketing pages keep this short on purpose. This page says more: what the foundation is, and what it makes possible.

The foundation: a governed memory

Facts are stored as records that keep their history. A correction is added as a new fact, never written over the old one, and every change carries an audit trail: what changed, when, and by whom. Connections into the tools a team already uses recall the right facts into a working session when they are needed, and write back what the session produces under the same discipline. The record is kept in open formats and runs in the customer's own cloud, on AWS, Azure, Google Cloud, or Cloudflare, with the customer holding the encryption keys.

What the foundation enables

A shared, governed memory is not only storage. Because every participant works through the same record, the record can carry the things agent work needs to be trustworthy inside a real company:

  • Agent identity. Agents take part as identified actors, not anonymous callers. Work is attributable: which agent, acting for whom, produced or changed what.
  • Agent communication. Agents exchange work through the shared record. A finding saved by one agent is there for the next, in a different tool, without side channels or copy and paste.
  • Collaboration. Several agents and people can work the same facts at the same time, each starting from what the others established. Approvals are part of the record, so a decision and its approval travel together.
  • Auditability. After the fact, the record answers what was known, what changed, and who approved it. That is what makes agent work governable, and reviewable, inside a company.

A harness for harnesses

Agents increasingly run inside harnesses: on a laptop, in a cloud, inside another platform. Each harness is good at running its own agents. None of them shares memory with the others, and an agent's work is usually trapped inside whichever harness ran it. Artifactum is the layer those harnesses have in common: whichever harness an agent runs in, it works from the same governed record. It does not replace the tools that run agents. It is the shared ground underneath them.

Across models and vendors

Artifactum is not tied to a single model or vendor. In daily use it runs across the major vendors: Anthropic, OpenAI, Meta (whose models include Llama, Wonder, and Spark), Google, and xAI, plus open source agents, all against the same facts. It also works with the personal agents people now use directly, Muse, Claude Duet, OpenAI Dots, and Grok Bot among them, because that is increasingly where the work happens: a person and their agent, drawing on the same governed record. Models will keep changing, and leads will change with them. The record stays.

Research

The thinking behind agent-to-agent communication over shared surfaces is written up in our whitepaper, Any Shared Surface Is a Channel, published in October 2026 under CC BY 4.0. It is vendor-neutral, and it is the baseline this platform is built on.

Where it runs, and how to start

A buildout is stood up in a standalone organization in the customer's own cloud, and the customer holds the keys. Artifactum is currently taking on a small number of early customers ahead of its enterprise launch, planned for Q2 2027. See early access, or read on in the docs for how the tested apps connect.