# Sezzlee Local LLM

> Sezzlee Local LLM is an open-source MCP server for ai models and llm tooling. MCP server that hands bounded language tasks to a local Ollama model from your agent. It runs locally on your machine and is started by your AI client.

- **Kind:** MCP server
- **Pricing:** Open source
- **Platforms:** CLI, macOS, Linux, Windows, self-hosted
- **Website:** https://docs.sezzlee.app/docs/llm-mcp/quickstart
- **Repository:** https://github.com/sezzlee/mcp/tree/HEAD/packages/servers/llm-mcp
- **MCP Registry name:** io.github.sezzlee/llm
- **Categories:** AI models and LLM tooling
- **Listing updated:** 2026-11-16

## About

The idea is to keep a planning agent's context small. The agent passes a file path, the server reads the file itself and runs the task on a model served by your own Ollama, and only the short answer returns to the agent. Files are read only inside a configured workspace folder, and output goes into a folder inside it.

Three tools are offered. local_task runs one summarize, extract, classify, transform or free task over short text or up to eight files, returning JSON when given a JSON Schema; summarize and extract chunk oversized input, other kinds refuse it, and nothing is silently truncated. local_map labels every row of a CSV (up to 2,000 rows, 2 to 50 labels) and writes a new file with a label column, returning counts and a few sample rows per label. local_status reports whether the model is reachable, its context window and input budget.

It needs Node.js 22 or later and Ollama running; the docs suggest qwen3:8b for an 8 GB GPU. Settings are passed as environment variables.

## Features

- **Local task runner:** Summarises, extracts, classifies or transforms text and files.
- **CSV labelling:** Labels each row of a CSV with chosen labels and writes a new file.
- **Status check:** Reports model reachability, context window and queue.
- **Private context:** File contents never enter the calling agent's context.
- **Workspace sandbox:** Reads only inside a configured folder.
- **Chunked long inputs:** Summarize and extract split oversized input without truncating.

## Good for

- Summarise meeting notes without sending them to a cloud model
- Tag thousands of support tickets by category
- Extract structured fields from files as JSON
- Offload routine text work from an expensive planning agent

## Install

### Claude Code (Terminal)

```sh
claude mcp add llm --env SEZZLEE_LLM_MODEL='<SEZZLEE_LLM_MODEL>' -- npx -y @sezzlee/llm-mcp
```

### Claude Desktop (claude_desktop_config.json)

```json
{
  "mcpServers": {
    "llm": {
      "command": "npx",
      "args": [
        "-y",
        "@sezzlee/llm-mcp"
      ],
      "env": {
        "SEZZLEE_LLM_MODEL": "<SEZZLEE_LLM_MODEL>"
      }
    }
  }
}
```

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

```json
{
  "mcpServers": {
    "llm": {
      "command": "npx",
      "args": [
        "-y",
        "@sezzlee/llm-mcp"
      ],
      "env": {
        "SEZZLEE_LLM_MODEL": "<SEZZLEE_LLM_MODEL>"
      }
    }
  }
}
```

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

```json
{
  "servers": {
    "llm": {
      "type": "stdio",
      "command": "npx",
      "args": [
        "-y",
        "@sezzlee/llm-mcp"
      ],
      "env": {
        "SEZZLEE_LLM_MODEL": "<SEZZLEE_LLM_MODEL>"
      }
    }
  }
}
```

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

```json
{
  "mcpServers": {
    "llm": {
      "command": "npx",
      "args": [
        "-y",
        "@sezzlee/llm-mcp"
      ],
      "env": {
        "SEZZLEE_LLM_MODEL": "<SEZZLEE_LLM_MODEL>"
      }
    }
  }
}
```

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

```json
{
  "mcpServers": {
    "llm": {
      "command": "npx",
      "args": [
        "-y",
        "@sezzlee/llm-mcp"
      ],
      "env": {
        "SEZZLEE_LLM_MODEL": "<SEZZLEE_LLM_MODEL>"
      }
    }
  }
}
```

## Tools

- `local_task`: Runs one bounded language task (classify, extract, summarize, transform or free) on the local model.
- `local_map`: Labels every row of a CSV file and writes a new file with a label column.
- `local_status`: Reports whether the local model is reachable, its context window, input budget and queue.

## FAQ

### What do I need installed?

Node.js 22 or later, jq for the docs recipes and a running Ollama with a model that can follow a JSON schema.

### Which model should I use?

The docs suggest qwen3:8b on a machine with 8 GB of GPU memory.

### What if Ollama is not running?

The server starts, local_status reports reachable false and tasks fail with backend_unavailable.

### Does it write files?

Only local_map, which adds new files to its own output directory.

### Is Sezzlee Local LLM free?

Yes. Sezzlee Local LLM is open source, so you can use and self-host it for free.

### How do I add Sezzlee Local LLM to Claude Code?

Run this command in your terminal: claude mcp add llm --env SEZZLEE_LLM_MODEL='<SEZZLEE_LLM_MODEL>' -- npx -y @sezzlee/llm-mcp

### What are the best alternatives to Sezzlee Local LLM?

The closest alternatives to Sezzlee Local LLM are Claude Code Ultimate Guide, signafAIde, Toolkien, Lugh and HypePaper.

## Alternatives

- [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.
- [signafAIde](https://ai-powered.com/mcp/signafaide): signafAIde is a freemium MCP server for security.
- [Toolkien](https://ai-powered.com/mcp/toolkien): Toolkien is an open-source MCP server for ai models and llm tooling.
- [Lugh](https://ai-powered.com/mcp/lugh): Lugh is an open-source MCP server for ai models and llm tooling.
- [HypePaper](https://ai-powered.com/mcp/hypepaper): HypePaper is a freemium MCP server for ai models and llm tooling.
- [Agent Data Tools: Compact (find, describe, run)](https://ai-powered.com/mcp/agent-data-tools-compact-find-describe-run): Agent Data Tools: Compact (find, describe, run) is a freemium MCP server for developer tools.
- [Angel Agent Guard: pre-purchase check + prompt-injection scan](https://ai-powered.com/mcp/angel-agent-guard-pre-purchase-check-prompt-injection-scan): Angel Agent Guard: pre-purchase check + prompt-injection scan is a freemium MCP server for ai models and llm tooling.
- [Mizan Calibration](https://ai-powered.com/mcp/mizan-calibration): Mizan Calibration is a free MCP server for ai models and llm tooling.

---
Source: https://ai-powered.com/mcp/sezzlee-local-llm
