Skip to content
Nappai Documentation

LM Studio Embeddings

LM Studio Embeddings lets you turn text into numerical vectors (embeddings) that can be used for searching, clustering, or feeding into other AI models. It connects to a local or remote LM Studio server, pulls the list of available models, and creates embeddings with a chosen model and temperature setting.

How it Works

When you add this component to your workflow, it first asks the LM Studio server for a list of available models. You pick one from the dropdown. The component then sends your text to the server’s embeddings endpoint, receives a vector representation, and outputs that vector. The whole process happens automatically, so you don’t need to write any code.

Inputs

Input Fields

  • LM Studio Base URL: The web address where your LM Studio server is running (e.g., http://localhost:1234/v1). This tells the component where to send requests.
  • Model: The specific LM Studio model you want to use for embeddings. The list is refreshed automatically from the server.
  • Model Temperature: A number that controls how random the model’s output can be. Lower values (e.g., 0.1) make the output more deterministic.

Outputs

  • Embeddings: A vector representation of the input text. You can feed this output into other components such as similarity search, clustering, or downstream AI models.

Usage Example

  1. Add the component to your dashboard and connect the text you want to embed to its input.
  2. Set the LM Studio Base URL to the address of your server (default is http://localhost:1234/v1).
  3. Choose a model from the dropdown. If you’re unsure, pick the first one listed.
  4. Leave the temperature at 0.1 for stable results.
  5. Run the workflow. The component will output an embeddings vector that you can use in the next step of your automation.
  • OpenAIEmbeddings – Uses OpenAI’s API to generate embeddings.
  • HuggingFaceEmbeddings – Generates embeddings from Hugging Face models.
  • VectorStore – Stores and searches embeddings in a database.

Tips and Best Practices

  • Check the server is running before running the workflow; otherwise the component will fail to retrieve models.
  • Use a specific model that matches your text domain for better quality embeddings.
  • Keep the temperature low (e.g., 0.1) unless you need more varied outputs.
  • Install langchain-nvidia-ai-endpoints if you haven’t already; the component won’t work without it.

Security Considerations

  • The LM Studio Base URL may expose sensitive endpoints. Use a secure network or VPN when connecting to a remote server.
  • The component currently uses a hard‑coded API key (1234). In production, replace this with a secure key or environment variable to protect your credentials.