Perceo Suite Documentation / Knowledge workspace

Overview

How Archivum stores human and team knowledge.

Archivum is the human knowledge layer of the Perceo Suite.

It is a self-hosted second brain for people and teams who want durable Markdown knowledge, calm editing, AI-native ingest, semantic search, graph exploration, and MCP access without giving up control of their data.

Info

Archivum is in development. The current positioning is intentionally narrow: human knowledge first, repo graph memory elsewhere.

ArchgraphSee how structured project memory stays separate from the human wiki.ArchductorSee how agent execution consumes memory without owning long-term knowledge.Computer-use testingSee where future behavioral verification fits in the stack.

Role in the suite

Archivum stores human knowledge.

Use it for notes, pages, folders, backlinks, daily notes, project notes, public sharing, and long-lived team memory. It is the place a human opens when they need to write, browse, curate, or explain what the team knows.

What it is

Server-hosted Obsidian-style workspace.
Markdown-first wiki with backlinks and graph navigation.
AI ingest for files, URLs, documents, media transcripts, and code.
Semantic search and question answering over local knowledge.
MCP server so agents can read, search, and write knowledge safely.

What it is not

It is not the repo, branch, commit, or issue graph. That belongs in Archgraph.
It is not the agent execution surface. That belongs in Archductor.
It is not the QA layer. That belongs in computer-use testing.

Current direction

Archivum is in development. The product priority is a sharp, self-hosted knowledge workspace that feels calm for humans and useful for agents.

The long-term win is compounding memory: every source you ingest becomes structured, searchable, linkable knowledge that future human and agent work can use.

bash

# planned local-first shape
docker compose up -d
open http://localhost:3000