Skill: text-embed

Version: 1.0.0

Generate embedding vectors from text using OpenAI or local models.

Perfect for RAG, semantic search, and similarity matching.

name: text-embed version: 1.0.0 description: Generate embedding vectors from text for RAG and semantic search

Credentials

credentials:

• name: OPENAI_API_KEY required: false description: OpenAI API key (not required if using Ollama)

Skill inputs

inputs:

• name: text type: string required: true description: Text to embed (up to 8191 tokens)
• name: model type: string default: text-embedding-3-small description: Embedding model to use

Skill outputs

outputs:

• name: embedding type: array description: Vector embedding (1536 or 3072 dimensions)
• name: metadata type: object description: Model info, dimensions, token count

Backend options

backends:

• name: openai type: remote requires: [OPENAI_API_KEY] description: OpenAI text-embedding-3-small (fast, ~$0.00002 per 1K tokens)
• name: ollama type: local models: [nomic-embed-text, all-minilm] description: Local Ollama embeddings (private, free)

Expanso components used

components: inputs: - stdin - http_server processors: - mapping - openai_embeddings - log outputs: - stdout - sync_response