# CEmbedding MCP server

> Converts text into vector embeddings using local ONNX models or OpenAI-compatible APIs.

- Repository: https://github.com/Cloto-dev/CEmbedding
- Page: https://hostedmcp.org/servers/cloto-dev/cembedding
- Transport: —
- Stars: 1
- License: MIT

## Environment variables

| Name | Required | Secret | Description |
|---|---|---|---|
| `CEMBEDDING_AUTH_TOKEN` | no | no |  |
| `CPERSONA_EMBEDDING_MODE` | no | no |  |
| `CPERSONA_EMBEDDING_URL` | no | no |  |
| `EMBEDDING_PROVIDER` | no | no |  |

## FAQ

### Can I run embeddings locally without an API key?

Yes, you can run models on-device via ONNX, which requires no API key and no network connection.

### How can I ensure two vectors are comparable?

You should store the fingerprint provided by the /capabilities endpoint; two vectors can be compared when their fingerprints are equal.
