RedisVL Documentation

What is RedisVL?

RedisVL is a comprehensive Go library for building AI-native applications with Redis. It provides a high-level interface for:

Vector similarity search with advanced filtering
Schema-based index management from YAML or programmatic configuration
Hybrid queries combining vector search with metadata filters and full-text scoring
Semantic and embedding caching for LLM applications
Local, in-process embedding models via ONNX Runtime — no API keys required
Flexible storage with both Redis Hash and JSON support

RedisVL for Golang is a port of the popular Python RedisVL library, bringing these capabilities to the Go ecosystem with idiomatic, context-aware APIs and goroutine-safe clients.

Resources

Getting Started

Learn the basics and create your first vector search application.

Get Started →

Querying

Combine vector search with powerful metadata filtering.

Learn More →

Semantic Caching

Cache LLM responses semantically to reduce costs and latency.

Explore Caching →

Vectorizers

Create embeddings with hosted APIs or local ONNX models.

Choose a Vectorizer →

Rerankers

Improve search quality with hosted or local cross-encoder models.

Boost Relevance →

MCP Server

Expose your indices to AI agents via the Model Context Protocol.

Connect Agents →

Features

RedisVL for Golang offers powerful features for AI-native applications:

Index Management - Design search schema and indices with ease from YAML or programmatic configuration
Advanced Vector Search - KNN, range, hybrid, and multi-vector queries with complex filtering support
Embedding Creation - OpenAI, Azure OpenAI, Cohere, Mistral, VoyageAI, Ollama, or local Hugging Face models via ONNX Runtime
Reranking - Improve search result quality with Cohere, VoyageAI, or local cross-encoder models
Semantic Caching - Cache LLM responses with semantic similarity, increasing QPS and decreasing system cost — self-hosted or via the managed LangCache service
Embeddings Cache - Cache vector embeddings to avoid redundant computation
Message History - Store chat history with recency- or relevancy-based retrieval
Semantic Routing - Route queries to topics and intents with vector similarity
Storage Flexibility - Choose between Hash and JSON storage based on your needs
Tooling - The rvl command-line interface and a built-in MCP server for AI agents

Installation

Add RedisVL to your project:

go get github.com/redis/redis-vl-golang@v0.2.0

For local embedding models (a separate module because it uses ONNX Runtime via cgo):

go get github.com/redis/redis-vl-golang/extensions/vectorize/hf@v0.2.0

Quick Example

package main

import (
    "context"
    "fmt"

    redisvl "github.com/redis/redis-vl-golang"
    "github.com/redis/redis-vl-golang/filter"
    "github.com/redis/redis-vl-golang/query"
    "github.com/redis/redis-vl-golang/schema"
)

func main() {
    ctx := context.Background()

    // Define the index schema
    vectorField, _ := schema.NewVectorField("embedding", schema.VectorAttrs{
        Dims: 384, Algorithm: schema.HNSW, Datatype: "float32",
        DistanceMetric: schema.Cosine,
    })
    s, _ := schema.NewIndexSchema(
        schema.IndexInfo{Name: "products", Prefixes: []string{"products"}},
        schema.NewTextField("description"),
        schema.NewTagField("category"),
        schema.NewNumericField("price"),
        vectorField,
    )

    // Create the index
    index, _ := redisvl.NewSearchIndexFromURL(s, "redis://localhost:6379")
    defer index.Close()
    index.Create(ctx)

    // Load data, then run a filtered vector search
    q := query.NewVectorQuery("embedding", queryVector).
        NumResults(5).
        Filter(filter.Tag("category").Eq("electronics").
            And(filter.Num("price").Le(100))).
        ReturnFields("description", "category", "price")

    results, _ := index.Query(ctx, q)
    fmt.Println(results)
}

Connecting with the Community

Join the Redis community to get help, share your experiences, and contribute: