
Quberi
Per-document permissions, so AI answers only draw on what each user may read
- Pricing
- Open source
- Platforms
- Web, self-hosted, API
- Runs
- Remotely
- Version
- 0.1.0
- Listing updated
- 2 days ago
About Quberi
Quberi is an open-source MCP server for knowledge and memory. It runs as a hosted remote server, so there is nothing to install locally.
Quberi's central idea is that access is decided per document rather than per space. Within one space, a deny rule on a document beats any allow rule or space role, retrieval filtering happens inside the vector index, and rules are checked on every request, so a change takes effect on the next one. The MCP server carries that model to outside agents: they search as the signed-in user and see exactly what that person could see.
The tools let an agent search, browse and save notes into the knowledge base. Organization admins have to switch on outside tool access, which is off by default and uses OAuth with PKCE, and the audit log records searches made by connected tools. Before a question reaches the chat model, the user can preview, edit or drop retrieved passages.
Quberi can be used as a hosted service or run yourself with Docker Compose under the AGPL-3.0 licence. Sign-in options include email code, Google and Feishu, embeddings run on CPU inside the API and worker, and chat models can be bring-your-own-key. The product pages are in Chinese, and the authors suggest manufacturing, energy and compliance-heavy companies as typical users.
Install
Run in a terminal
claude mcp add --transport http mcp https://mcp.quberi.com/Add to Settings → Connectors → Add custom connector
https://mcp.quberi.com/Add to ~/.cursor/mcp.json
{
"mcpServers": {
"mcp": {
"url": "https://mcp.quberi.com/"
}
}
}Add to .vscode/mcp.json
{
"servers": {
"mcp": {
"type": "http",
"url": "https://mcp.quberi.com/"
}
}
}Add to ~/.codeium/windsurf/mcp_config.json
{
"mcpServers": {
"mcp": {
"serverUrl": "https://mcp.quberi.com/"
}
}
}Add to your client's MCP settings (most use this shape)
{
"mcpServers": {
"mcp": {
"type": "http",
"url": "https://mcp.quberi.com/"
}
}
}What it does
- Per-document permissions
- Deny rules win, and users in one space can see different documents.
- Permission-filtered retrieval
- Access filtering happens inside the vector index rather than after results return.
- Search, browse and save notes
- The MCP server reads and adds to the knowledge base as the signed-in user.
- Audit log
- Records sensitive actions, including searches from connected tools.
- Send review
- Preview, edit or drop passages before they go to the chat model.
- Self-hostable
- Runs with Docker Compose under AGPL-3.0, or use the hosted service.
Good for
- Let an external agent search company documents without widening access
- Keep customer drawings, quotes and contracts visible only to the right people
- Audit which documents a model saw for a question
- Save a note from an agent into a team knowledge base
Pricing
Open source
AGPL-3.0 code; hosted service pricing not published
Prices change. Check ai.quberi.com before you buy.
Quberi FAQ
Who can the agent see documents as?
As the signed-in user, so it finds only what that person is allowed to read.
Is outside access on by default?
No. An organization admin has to allow read-only search from outside tools.
Can I run it myself?
Yes, with Docker Compose. The code is licensed under AGPL-3.0.
How do users sign in?
By email code, Google or Feishu.
Is Quberi free?
Yes. Quberi is open source, so you can use and self-host it for free.
How do I add Quberi to Claude Code?
Run this command in your terminal: claude mcp add --transport http mcp https://mcp.quberi.com/
What are the best alternatives to Quberi?
The closest alternatives to Quberi are Codebase Memory, XERJ, Claude Code Ultimate Guide, Nautilus Compass Memory and note1.
Alternatives to Quberi

Codebase Memory
MCP serverLocal MCP server that builds a code knowledge graph for AI coding agents
Runs locally- Open source
- 48.7K GitHub stars
- Knowledge and memory

XERJ
MCP serverLocal search engine for AI agents with cited answers over your code and docs
Runs locally- Open source
- 19.2K GitHub stars
- Knowledge and memory

Claude Code Ultimate Guide
MCP serverMCP server for searching the Claude Code Ultimate Guide from any MCP client
Runs locally- Open source
- 6.5K GitHub stars
- AI models and LLM tooling

Nautilus Compass Memory
MCP serverLocal-first long-term memory layer for AI agents, served over MCP
Not checked yet- 5 tools
- Remote
- Open source
- 1.3K GitHub stars
- Knowledge and memory

note1
MCP serverMCP server giving AI assistants search and summaries across your note1 meetings
- 4 tools
- Remote
- Free trial
- 1 GitHub star
- Productivity
MedXpert 图书馆连接器(通用版)
MCP serverLocal MCP server that searches and answers questions over your own Markdown wiki
Runs locally- 4 tools
- Open source
- Knowledge and memory