# Nefesh — Real-Time Human State Awareness for AI

Nefesh gives AI agents a live human-state signal, a stress score, and a suggested response based on incoming biometric and text data.

- MCP server · `type=mcp` · Claude Code, Codex, Cursor, GitHub Copilot, Windsurf (Devin Desktop), Cline, Zed, Gemini CLI
- Author: nefesh-ai
- License: MIT · policy=link-only · description only; the file stays with the author
- Source: https://github.com/nefesh-ai/nefesh-mcp-server
- Updated: 2026-04-16
- HTML: https://affpapa.org/en/nevibecode/mcp/nefesh-real-time-human-state-awareness-for-ai

## What it does

Nefesh accepts signals such as heart rate, voice characteristics, facial expression, and text sentiment, then returns one combined human-state result. The response includes a stress score from 0 to 100 and a suggested action such as maintaining the current approach, simplifying it, de-escalating, or pausing.

Its MCP tools cover state lookup, signal ingestion, trigger memory, and timestamped session history. From the second call in a session onward, the result can show `adaptation_effectiveness`, so an agent can tell whether its previous response reduced stress. Connections use Streamable HTTP, with no local package to install. Authenticated tools require an `X-Nefesh-Key` value.

Nefesh also exposes an A2A endpoint at `https://mcp.nefesh.ai/a2a`, allowing other agents to communicate with it as a peer. A key can be requested through the agent after confirming an email link, or added to the MCP configuration before connecting.

## Requirements
- Remote server: `https://mcp.nefesh.ai/mcp`

## Install

### Claude Code (native platform)

**Platform CLI: remote server**
```bash
claude mcp add --transport http human-state https://mcp.nefesh.ai/mcp
```

### Codex

**Platform CLI: remote server**
```bash
codex mcp add nefesh-real-time-human-state-awareness-for-ai --url https://mcp.nefesh.ai/mcp
```

### Cursor

**Config file**
File `.cursor/mcp.json`:
```json
{
    "mcpServers": {
        "nefesh-real-time-human-state-awareness-for-ai": {
            "url": "https://mcp.nefesh.ai/mcp"
        }
    }
}
```

### GitHub Copilot

**Config file**
File `.vscode/mcp.json`:
```json
{
    "servers": {
        "nefesh-real-time-human-state-awareness-for-ai": {
            "type": "http",
            "url": "https://mcp.nefesh.ai/mcp"
        }
    }
}
```

### Gemini CLI

**Platform CLI: remote server**
```bash
gemini mcp add -t http nefesh-real-time-human-state-awareness-for-ai https://mcp.nefesh.ai/mcp
```

## When to use it
- read a person’s current state from sensor signals
- turn several inputs into one stress assessment
- choose whether the agent should simplify or pause
- measure whether the previous response helped
- inspect session history and its stress trend

## Example prompts
- Check the person’s current state and choose a response that should lower the tension.
- Send heart rate, voice tone, and text sentiment to Nefesh with `ingest`, then explain the returned state.
- Review the session history and tell me whether stress is rising or falling.

## Similar
- [real-time-features](https://affpapa.org/en/nevibecode/skills/real-time-features.md) — skill · 543 installs
- [dashboards-and-real-time-visualization](https://affpapa.org/en/nevibecode/skills/dashboards-and-real-time-visualization.md) — skill · 29 installs
- [real-estate-expert](https://affpapa.org/en/nevibecode/skills/real-estate-expert.md) — skill · 820 installs
- [cli-a2a](https://affpapa.org/en/nevibecode/skills/cli-a2a.md) — skill · 1000 installs
- [google-agents-cli-publish](https://affpapa.org/en/nevibecode/skills/google-agents-cli-publish.md) — skill · 511900 installs
- [langgraph-human-in-the-loop](https://affpapa.org/en/nevibecode/skills/langgraph-human-in-the-loop.md) — skill · 17200 installs

---
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