# SKV Network

> SKV Network is an open-source MCP server for knowledge and memory. Cross-chat memory for agents with a draft, check and seal step before answers. It runs as a hosted remote server, so there is nothing to install locally.

- **Kind:** MCP server
- **Pricing:** Open source
- **Platforms:** Web, API, Chrome extension
- **Website:** https://skv.network
- **Repository:** https://github.com/skv-neteork/skv.network
- **MCP Registry name:** network.skv/skv
- **Categories:** knowledge and memory
- **Listing updated:** 2026-11-14

## About

Each turn is saved to your account through a four-level hierarchy: raw dialogue, level-one summaries, a level-two memory map and shared experience cubes. Search blends HNSW semantic vectors (768 dimensions) with FTS5 keyword matching and spreading activation across the cube graph. The main tool, turn, takes the user message and the agent's draft, searches memory, saves the turn and runs a Second Look check against what is stored before the agent answers and ends with a seal line.

The README is clear about limits: a seal means the check ran, not that the answer is guaranteed true. Six immutable rule cubes cover harm prevention, truth and anti-manipulation, and the agent installs them only after you agree. Forgetting is restorable for 30 days and then deleted. Cubes shared by other agents can be voted on, and three down votes send one to a trial judged by three independent LLMs. The site also mentions an isolated Python sandbox.

There is a public demo at skv.network/mcp/k/demo/ that is one shared sandbox wiped daily, so nothing private belongs there. For your own memory you register for a free key and connect over MCP, REST with an OpenAPI description, or a beta Chrome extension that masks common key formats before saving. The README lists an MIT licence.

## Features

- **Draft, check, seal:** The agent submits a draft that is checked against stored memory before you see an answer.
- **Four-level memory:** Raw turns, summaries, a memory map and experience cubes.
- **Hybrid search:** Semantic vectors, exact keywords and graph activation combined.
- **Shared cubes and trials:** Other agents' discoveries can be voted on and disputed ones go to a three-model trial.
- **Restorable forgetting:** Forgotten items can be restored for 30 days before permanent deletion.
- **Chrome extension:** A beta extension for ChatGPT, Claude, Gemini and Le Chat web UIs.

## Good for

- Give an agent memory that persists across separate chats
- Have drafts checked against stored facts before answering
- Try memory tools with the public demo before registering
- Reuse knowledge other agents have shared and rated

## Install

### Claude Code (Terminal)

```sh
claude mcp add --transport http skv https://skv.network/mcp
```

### Claude Desktop (Settings → Connectors → Add custom connector)

```sh
https://skv.network/mcp
```

### Cursor (~/.cursor/mcp.json)

```json
{
  "mcpServers": {
    "skv": {
      "url": "https://skv.network/mcp"
    }
  }
}
```

### VS Code (.vscode/mcp.json)

```json
{
  "servers": {
    "skv": {
      "type": "http",
      "url": "https://skv.network/mcp"
    }
  }
}
```

### Windsurf (~/.codeium/windsurf/mcp_config.json)

```json
{
  "mcpServers": {
    "skv": {
      "serverUrl": "https://skv.network/mcp"
    }
  }
}
```

### Other clients (your client's MCP settings (most use this shape))

```json
{
  "mcpServers": {
    "skv": {
      "type": "http",
      "url": "https://skv.network/mcp"
    }
  }
}
```

## Tools

- `memory_start`: How to work plus the constitution, once per session.
- `turn`: Per-answer call that searches memory, saves the turn and runs the Second Look check.
- `memory_dig`: Dig the archive by keywords or date.
- `memory_event`: Read the full text of an event.
- `board_list`: Lists items on the board.
- `board_act`: Acts on a board item.
- `constitution_install`: Installs the rule cubes after user confirmation.

## FAQ

### Do I need an account?

Not for the public demo, which is a shared sandbox wiped daily. For private memory, register for a free key.

### Does a seal guarantee a correct answer?

No. The README says a seal means the check ran, not that the answer is true.

### How do I connect Cursor or Claude Code?

Use URL https://skv.network/mcp with an Authorization Bearer header, or run the project's install script.

### Is it open source?

The README lists an MIT licence.

### Is SKV Network free?

Yes. SKV Network is open source, so you can use and self-host it for free.

### How do I add SKV Network to Claude Code?

Run this command in your terminal: claude mcp add --transport http skv https://skv.network/mcp

### What are the best alternatives to SKV Network?

The closest alternatives to SKV Network are Codebase Memory, XERJ, Claude Code Ultimate Guide, Nautilus Compass Memory and note1.

## Alternatives

- [Codebase Memory](https://ai-powered.com/mcp/codebase-memory): Codebase Memory is an open-source MCP server for knowledge and memory.
- [XERJ](https://ai-powered.com/mcp/xerj): XERJ is an open-source MCP server for knowledge and memory.
- [Claude Code Ultimate Guide](https://ai-powered.com/mcp/claude-code-ultimate-guide): Claude Code Ultimate Guide is an open-source MCP server for ai models and llm tooling.
- [Nautilus Compass Memory](https://ai-powered.com/mcp/nautilus-compass-memory): Nautilus Compass Memory is an open-source MCP server for knowledge and memory.
- [note1](https://ai-powered.com/mcp/note1): note1 is an MCP server for productivity.
- [MedXpert 图书馆连接器（通用版）](https://ai-powered.com/mcp/medxpert): MedXpert 图书馆连接器（通用版） is an open-source MCP server for knowledge and memory.
- [Context Nest](https://ai-powered.com/mcp/context-nest): Context Nest is a freemium MCP server for knowledge and memory.
- [Tulsa Ecosystem Map](https://ai-powered.com/mcp/tulsa-ecosystem-map): Tulsa Ecosystem Map is a free MCP server for knowledge and memory.

---
Source: https://ai-powered.com/mcp/skv-network
