Redis Stack¶

Type: redis

Uses Redis Stack (the redis/redis-stack Docker image) with the RediSearch module for vector similarity search via FT.CREATE / FT.SEARCH.

Redis Stack required

Standard Redis does not include the FT commands. Use the redis/redis-stack or redis/redis-stack-server image, or enable the redisearch module manually.

Docker quickstart¶

docker run -d -p 6379:6379 redis/redis-stack-server

Configuration¶

- name: embedder
  type: embedding/openai
  metadata:
    base_url: http://localhost:11434/v1
    model: nomic-embed-text
    dimensions: "768"

- name: redis-docs
  type: redis
  metadata:
    addr: localhost:6379
    index: daimon-docs
    embedder: embedder
    dimensions: "768"

Metadata keys¶

Key	Default	Description
`addr`	`localhost:6379`	Redis address
`password`	—	Redis password
`db`	`"0"`	Redis database number
`index`	`daimon`	RediSearch index name
`embedder`	—	Name of a declared `embedding/openai` component
`dimensions`	`"1536"`	Vector dimensions (must match the embedder)

Notes¶

The FT index is created automatically on startup if it does not exist (HNSW, FLOAT32, cosine distance).
Each document is stored as a Redis Hash keyed {index}:{id}.
dimensions must match the embedder's output size exactly. Mismatches produce a Redis error at query time.