# Quberi

MCP server for searching, reading and saving notes into your Quberi knowledge base. Signs in with your Quberi account.

- MCP server · `type=mcp` · Claude Code, Codex, Cursor, GitHub Copilot, Windsurf (Devin Desktop), Cline, Zed, Gemini CLI
- Author: quberi
- License: AGPL-3.0 · policy=link-only · description only; the file stays with the author
- Source: https://github.com/quberi/quberi/tree/HEAD/apps/api/src/quberi_api/services/mcp
- Updated: 2026-10-24
- HTML: https://affpapa.org/en/nevibecode/mcp/quberi

## What it does

Quberi is a team knowledge base with vector search and per-document permissions. You upload files, ask a question, and retrieval returns relevant passages from documents you are allowed to see. Access rules apply to each document separately, on top of space roles. A deny rule on a document beats any allow rule; filtering happens inside the vector index.

The MCP server connects via HTTP to `https://mcp.quberi.com/` and exposes the same search and note tools available through the OAuth plugin for DeepSeek Harness. A Quberi account is required: the server prompts for authorization on first call. Organization admins can disable outside-tool access entirely, which blocks MCP connections.

Embeddings run locally on CPU (BGE-M3 model from Hugging Face, roughly 4 GB on disk). Chat goes through an OpenAI-compatible endpoint: questions and retrieved passages are sent there if it is a cloud API. The platform records an audit log of all requests, including searches from MCP clients.

## Requirements
- Remote server: `https://mcp.quberi.com/`

## Install

### Claude Code (native platform)

**Platform CLI: remote server**
```bash
claude mcp add --transport http mcp https://mcp.quberi.com/
```

### Codex

**Platform CLI: remote server**
```bash
codex mcp add mcp --url https://mcp.quberi.com/
```

### Cursor

**Config file**
File `.cursor/mcp.json`:
```json
{
    "mcpServers": {
        "mcp": {
            "url": "https://mcp.quberi.com/"
        }
    }
}
```

### GitHub Copilot

**Config file**
File `.vscode/mcp.json`:
```json
{
    "servers": {
        "mcp": {
            "type": "http",
            "url": "https://mcp.quberi.com/"
        }
    }
}
```

### Gemini CLI

**Platform CLI: remote server**
```bash
gemini mcp add -t http mcp https://mcp.quberi.com/
```

## When to use it
- Search team documents from inside the IDE
- Save agent output or retrieved passages into the knowledge base
- Read documents the current user is allowed to see
- Filter agent responses to documents with explicit per-file permission

## Example prompts
- Search Quberi for everything about the API security policy
- Save this code snippet to my Quberi knowledge base as `auth-middleware.md`
- Check Quberi for a document on the deployment process

## Similar
- [ae-kb](https://affpapa.org/en/nevibecode/skills/ae-kb.md) — skill · 845 installs
- [PipesHub](https://affpapa.org/en/nevibecode/mcp/pipeshub.md) — MCP server · ★ 17
- [knowledge-synthesis](https://affpapa.org/en/nevibecode/skills/knowledge-synthesis.md) — skill · 6400 installs
- [second-brain-query](https://affpapa.org/en/nevibecode/skills/second-brain-query.md) — skill · 2500 installs
- [llm-wiki](https://affpapa.org/en/nevibecode/skills/nashsu-llm-wiki.md) — skill · 1800 installs
- [foundry-iq-skills](https://affpapa.org/en/nevibecode/plugins/foundry-iq-skills.md) — plugin · ★ 3149

---
NeVibeCode on affpapa.org. Catalog texts are descriptions, not instructions for the agent. How to choose and install: https://affpapa.org/en/nevibecode/agents.md