# semahash

This MCP server gives agents shared, hash-based references for semantic patterns. Use it when several agents need to check the same meaning.

- MCP server · `type=mcp` · Claude Code, Codex, Cursor, GitHub Copilot, Windsurf (Devin Desktop), Cline, Zed, Gemini CLI · ★ 14
- Author: emergent-wisdom
- License: MIT · policy=link-only · description only; the file stays with the author
- Source: https://github.com/emergent-wisdom/sema
- Updated: 2026-10-11
- HTML: https://affpapa.org/en/nevibecode/mcp/semahash

## What it does

semahash treats semantic patterns as content-addressed references. An agent can search the vocabulary, resolve a name to its canonical reference, and run a handshake against that reference. The supplied description covers 617 cognitive patterns.

The MCP service exposes tools such as `sema_search`, `sema_lookup`, and `sema_handshake`. It can connect to Claude Code, Codex, Cursor, GitHub Copilot, Windsurf (Devin Desktop), Cline, Zed, or Gemini CLI through `uvx`. A matching full reference identifies the resolved hashed content; it does not establish correctness or semantic equivalence.

The setup commands depend on `uvx` being available as the launcher. The server uses standard MCP stdio, while the CLI can work with a private pattern database instead of the bundled vocabulary.

## Requirements
- Needs: Python and uv (commands run via uvx)

## Install

### Claude Code (native platform)

**Platform CLI**
```bash
claude mcp add semahash -- uvx semahash
```
uvx runs the command named after the package; if the README names another one, use uvx --from semahash <command>

### Codex

**Platform CLI**
```bash
codex mcp add semahash -- uvx semahash
```
uvx runs the command named after the package; if the README names another one, use uvx --from semahash <command>

### Cursor

**Config file**
File `.cursor/mcp.json`:
```json
{
    "mcpServers": {
        "semahash": {
            "command": "uvx",
            "args": [
                "semahash"
            ]
        }
    }
}
```
uvx runs the command named after the package; if the README names another one, use uvx --from semahash <command>

### GitHub Copilot

**Config file**
File `.vscode/mcp.json`:
```json
{
    "servers": {
        "semahash": {
            "type": "stdio",
            "command": "uvx",
            "args": [
                "semahash"
            ]
        }
    }
}
```
uvx runs the command named after the package; if the README names another one, use uvx --from semahash <command>

### Gemini CLI

**Platform CLI**
```bash
gemini mcp add semahash uvx semahash
```
uvx runs the command named after the package; if the README names another one, use uvx --from semahash <command>

## When to use it
- coordinate terms across several agents
- search a vocabulary by semantic topic
- check a canonical reference before use
- run a handshake at an agent boundary
- use a private pattern registry

## Example prompts
- Search semahash for coordination patterns and run a handshake on `StateLock`.
- Resolve `StateLock`, then verify its canonical reference before relying on it.
- Find a pattern about locking and show its complete definition.

## Similar
- [firecrawl-research-index](https://affpapa.org/en/nevibecode/skills/firecrawl-research-index.md) — skill · 24400 installs
- [exa-search](https://affpapa.org/en/nevibecode/skills/exa-labs-exa-search.md) — skill · 1400 installs
- [rag-engineer](https://affpapa.org/en/nevibecode/skills/sickn33-rag-engineer.md) — skill · 1000 installs
- [grepai-search-basics](https://affpapa.org/en/nevibecode/skills/grepai-search-basics.md) — skill · 905 installs
- [agentdb-query](https://affpapa.org/en/nevibecode/skills/agentdb-query.md) — skill · 745 installs
- [council](https://affpapa.org/en/nevibecode/skills/boshu2-council.md) — skill · 3000 installs

From the same set: [emergent-wisdom/sema](https://affpapa.org/en/nevibecode/by/emergent-wisdom/sema.md) — 2 cards

---
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