# onet-match

> onet-match is an open-source MCP server for hiring and recruiting. Map free text to ranked O*NET-SOC occupation codes, with tasks, skills and interests. It runs locally on your machine and is started by your AI client.

- **Kind:** MCP server
- **Pricing:** Open source
- **Platforms:** CLI, self-hosted, API
- **Repository:** https://github.com/benjamin-v-chan/onet-match
- **MCP Registry name:** io.github.Benjamin-V-Chan/onet-match
- **Categories:** [hiring and recruiting](https://ai-powered.com/categories/hiring-and-recruiting), [AI models and LLM tooling](https://ai-powered.com/categories/ai-and-llm)
- **Listing updated:** 2026-11-18

## About

The classifier turns messy text such as job titles, resume bullets, postings or interests into standardised occupation codes using a fine-tuned bge-base retriever and a cross-encoder reranker over all 923 O*NET 31.0 occupations. The README reports 51.1% top-1 on held-out O*NET text against 20.5% for a TF-IDF baseline, and 65.4% top-1 on messy human-style text, with calibrated confidence (ECE 0.037) and about 47 ms per query on a laptop.

Each match carries SOC code, title, confidence, major group, job zone, RIASEC interest code, description and top tasks, skills and knowledge with O*NET importance and level, plus a profile aggregated across matches. The MCP server runs locally over stdio and offers two read-only tools, match_occupations and get_occupation. A remote streamable HTTP mode is available too.

The first call downloads a model bundle of about 0.5 GB from Hugging Face. The same package ships a Python API, CLI, FastAPI REST service and Docker setup.

## Features

- **Ranked occupation matches:** Free text goes to SOC codes with calibrated confidence.
- **Enriched results:** Tasks, skills, knowledge and RIASEC interests per match.
- **Runs locally:** No API key or account; the model runs on your machine.
- **Several interfaces:** Python, CLI, REST API, MCP and Docker.
- **Reported accuracy:** 51.1% top-1 on O*NET test text and 65.4% on messy text.

## Good for

- Normalise informal job titles in postings
- Code resume bullets to occupations
- Suggest careers from a student's interests
- Batch-code labour-market text for analytics

## Install

### Claude Code (Terminal)

```sh
claude mcp add onet-match -- uvx onet-match
```

### Claude Desktop (claude_desktop_config.json)

```json
{
  "mcpServers": {
    "onet-match": {
      "command": "uvx",
      "args": [
        "onet-match"
      ]
    }
  }
}
```

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

```json
{
  "mcpServers": {
    "onet-match": {
      "command": "uvx",
      "args": [
        "onet-match"
      ]
    }
  }
}
```

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

```json
{
  "servers": {
    "onet-match": {
      "type": "stdio",
      "command": "uvx",
      "args": [
        "onet-match"
      ]
    }
  }
}
```

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

```json
{
  "mcpServers": {
    "onet-match": {
      "command": "uvx",
      "args": [
        "onet-match"
      ]
    }
  }
}
```

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

```json
{
  "mcpServers": {
    "onet-match": {
      "command": "uvx",
      "args": [
        "onet-match"
      ]
    }
  }
}
```

## Tools

- `match_occupations`: Match text to ranked O*NET-SOC occupations (text, top_k)
- `get_occupation`: Get details for one occupation by SOC code

## FAQ

### Does it need an API key?

No. It runs locally and is MIT licensed.

### How accurate is it?

The README reports 51.1% top-1 and 79.1% top-5 on held-out O*NET text, and 65.4% and 92.9% on messy text.

### How big is the download?

The first call fetches a model bundle of about 0.5 GB.

### What MCP tools exist?

match_occupations and get_occupation, both read-only.

### Is onet-match free?

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

### How do I add onet-match to Claude Code?

Run this command in your terminal: claude mcp add onet-match -- uvx onet-match

### What are the best alternatives to onet-match?

The closest alternatives to onet-match are headroom, omniget, next-ai-draw-io, mcp-for-blender and Claude Code Ultimate Guide.

## Where this comes from

- Description: written from the project's README.
- Install steps: from the official MCP Registry entry.

## Alternatives

- [headroom](https://ai-powered.com/mcp/headroom-3): headroom is an open-source MCP server for ai models and llm tooling.
- [omniget](https://ai-powered.com/mcp/omniget): omniget is an open-source MCP server for ai models and llm tooling.
- [next-ai-draw-io](https://ai-powered.com/mcp/next-ai-draw-io-2): next-ai-draw-io is an open-source MCP server for ai models and llm tooling.
- [mcp-for-blender](https://ai-powered.com/mcp/mcp-for-blender): mcp-for-blender is an open-source MCP server for design and media.
- [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.
- [unity-mcp](https://ai-powered.com/mcp/unity-mcp-2): unity-mcp is an open-source MCP server for ai models and llm tooling.
- [fastapi_mcp](https://ai-powered.com/mcp/fastapi-mcp): fastapi_mcp is an open-source MCP server for ai models and llm tooling.
- [tyk](https://ai-powered.com/mcp/tyk): tyk is an open-source MCP server for developer tools.

---
Source: https://ai-powered.com/mcp/onet-match
