SQLite Adapter

The SQLite adapter stores memories in a local SQLite database using better-sqlite3. It’s perfect for development, prototyping, and applications that don’t need a separate database server.

Installation

npm install @youcraft/recall-adapter-sqlite

Usage

import { createMemory } from '@youcraft/recall'
import { sqliteAdapter } from '@youcraft/recall-adapter-sqlite'

const db = sqliteAdapter({
  filename: 'recall.db',
})

const memory = createMemory({ db, embeddings, extractor })

Configuration

Option	Type	Default	Description
filename	string	"recall.db"	Path to the SQLite database file. Use ":memory:" for an in-memory database.

Examples

File-based database

const db = sqliteAdapter({
  filename: './data/memories.db',
})

In-memory database

Useful for testing or ephemeral sessions:

const db = sqliteAdapter({
  filename: ':memory:',
})

How It Works

The adapter automatically:

• Creates the memories table if it doesn’t exist
• Uses WAL mode for better concurrent read performance
• Stores embeddings as JSON strings
• Performs cosine similarity search in JavaScript (suitable for small-to-medium datasets)

Schema

The adapter creates this table structure:

CREATE TABLE IF NOT EXISTS memories (
  id TEXT PRIMARY KEY,
  user_id TEXT NOT NULL,
  content TEXT NOT NULL,
  embedding TEXT NOT NULL,
  metadata TEXT NOT NULL DEFAULT '{}',
  created_at TEXT NOT NULL,
  updated_at TEXT NOT NULL
);
CREATE INDEX IF NOT EXISTS idx_memories_user_id ON memories(user_id);

Performance Considerations

• Vector search: Performed in JavaScript using cosine similarity. Works well for up to ~10,000 memories per user.
• For larger datasets: Consider PostgreSQL with pgvector for native vector operations.
• WAL mode: Enabled by default for better read concurrency.