Contorium exposes structured project facts, decisions and handoff context through a local MCP bridge for coding assistants.
Contorium exposes structured project facts, decisions and handoff context through a local MCP bridge for coding assistants.
Contorium stores project intelligence such as decisions, constraints, risks and validation context. The MCP bridge supports capture, inspection and transfer into an assistant's current conversation. It does not execute coding tasks on its own; the host agent remains responsible for the requested implementation.
The documented @contorium/mcp package runs as a local process. Connect it from the intended project and initialize or synchronize the project's state. The bridge's optional AI layer is separate from the core structured store, so an LLM service is not required for every local operation.
transfer_project exports context in several modes, including handoff and intelligence. Older transfer names remain compatibility aliases. Start with one recorded decision, inspect its source and validation state, then export a concise handoff for a new session. A decision can become stale as architecture changes; update its evidence and retain the distinction between an accepted decision and a proposal awaiting confirmation.
Check the linked publisher documentation and your connected client's tool list. Names and availability can vary with the server version and account permissions.
| Tool | What it does |
|---|---|
| sessions | tools |
| transfer_project | Export project intelligence in the selected context mode. |
{
"mcpServers": {
"contorium": {
"command": "npx",
"args": [
"@contorium/mcp"
]
}
}
}Node.js 18 or later, the Contorium package and the intended initialized project workspace.
Store entities, relations and observations in a local knowledge graph with the reference Memory MCP server.
A structured scratchpad for hard problems — stepwise reasoning with revisions, branches and visible logic.
> **real long-term memory for ai agents. not rag. not a vector db. self-hosted.**
Phaedo carries a local preference projection into assistants and offers consultation tools at selected decision points.
It is a open-source universal memory engine where users can store and retrieve their AI conversations and context across different models. Users can
Local-first knowledge graph that watches your project files, extracts entities and relationships via LLMs, and lets you query across projects in