Ligantic

/

/

/

/

AI - Vector Context Query

Type identifier: ai:vectorContext:query Category: AI Operations

Description

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.

Input Handles

Handle	Type	Description
`query`	`string`	The text query to search for.
`<labelFilter>`	`string`	Dynamic handles for label filter values.

Output Handles

Handle	Type	Description
`results`	`array`	Array of matching results with content and scores.

Configuration Options

Option	Type	Default	Description
Top Count	`number`	`3`	Maximum number of results to return.
Label Filters	`string[]`	`[]`	Label keys to filter results by.
Label Values	`string[]`	`[]`	Expected label values (if static).

Behaviour

Receives query text from input handle
Computes embedding for the query
Searches vector context for similar embeddings:
- Applies label filters if configured
- Ranks by cosine similarity
- Limits to top N results
Returns matching content with similarity scores

Similarity Scoring

Results are ranked by cosine similarity:

Higher scores indicate more semantic similarity
Scores range from 0 to 1
Results are returned in descending score order

Label Filtering

When label filters are configured:

Only embeddings matching all specified labels are considered
Dynamic handles provide label values at runtime
Multiple filters create AND conditions

Result Structure

Each result contains:

id: The embedding's unique identifier
content: The original stored text
score: Similarity score (0-1)
Labels: All stored label values

Examples

Basic Semantic Search

Configuration:

Top Count: 5
Label Filters: (none)

Input:

"How do I reset my password?"

Output: Top 5 most semantically similar stored content items.

Filtered Search

Configuration:

Top Count: 3
Label Filters: ["documentId"]

Use case: Search within a specific document's chunks.

RAG (Retrieval-Augmented Generation)

Flow:

User asks a question
AI - Vector Context Query finds relevant content
Text - Template combines question + context
AI - Chat Message generates informed response

Configuration:

Top Count: 5
Provides relevant context to the LLM

Multi-Filter Search

Configuration:

Label Filters: ["category", "language"]

Use case: Search documentation by topic and language.

Related pages

AI - Chat Message

The AI - Chat Message node enables conversations with large language models (LLMs) from multiple providers including OpenAI, Anthropic, Google AI, and Groq. It supports dynamic message templates with variable interpolation, streaming responses, extended thinking, and AI actions (tool use).

AI - Compute Embedding

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.

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 Remove

The AI - Vector Context Remove node deletes embeddings from a vector context. It can remove specific embeddings by ID or remove all embeddings matching certain label criteria.

Text - Template

Renders a rich-content template with variable substitution. Insert variables through the template editor, then connect each generated input handle to the value it should render. It generates dynamic content such as messages, emails, or prompts.