AI - Compute Embedding

Type identifier: ai:llm:computeEmbedding Category: AI Operations

Description

The AI - Compute Embedding node converts text into high-dimensional vector representations (embeddings) that capture semantic meaning. These embeddings can be used for semantic search, similarity comparison, and vector context operations.

Input Handles

Handle	Type	Description
`text`	`string`	The text to compute embeddings for.
`<label>`	varies	Dynamic handles for label values.

Output Handles

Handle	Type	Description
`embedding`	`number[]`	The computed embedding vector.

Configuration Options

Option	Type	Default	Description
Labels	`string[]`	`[]`	Labels for categorising embeddings.

Behaviour

Receives text input from the input handle
Resolves any label values from dynamic handles
Sends text to embedding model (OpenAI text-embedding-ada-002)
Returns the embedding vector

Embedding Models

Currently supported:

OpenAI text-embedding-ada-002: 1536-dimensional vectors
Cohere embed-multilingual-v3: 1024-dimensional vectors
Cohere embed-english-v3: 1024-dimensional vectors

For the canonical model list, see AI - Model Reference — Compute Embedding Models.

Labels

Labels provide metadata for embeddings stored in vector contexts:

Used for filtering in vector queries
Each label creates a dynamic input handle for its value

Examples

Basic Embedding

Configuration:

Labels: (empty)

Input:

"The quick brown fox jumps over the lazy dog."

Output (embedding):

[0.0023, -0.0089, 0.0145, ...]  // 1536 dimensions

Labelled Embedding for Vector Storage

Configuration:

Labels: ["category", "source"]

Dynamic inputs:

category: "documentation"
source: "user-manual"

Use case: Compute embeddings with metadata for organised vector storage.

AI - Vector Context Add

The AI - Vector Context Add node stores text embeddings in a vector context, enabling later semantic search operations. It automatically computes embeddings and indexes them with optional labels for filtering.

AI - Vector Context Query

The AI - Vector Context Query node performs semantic search over stored embeddings in a vector context. It returns the most similar content based on embedding distance, with optional filtering by labels.

Text - Recursive Split

The Text - Recursive Split node divides text into smaller chunks suitable for processing by language models or other text analysis tools. It splits on natural boundaries (paragraphs, lines, words) where possible, while respecting the configured chunk size. Overlapping between chunks ensures context is preserved across boundaries.