createMnemo()

Create a Mnemo memory instance.

Signature

function createMnemo(config: MnemoConfig): Promise<MnemoInstance>

Examples

import { createMnemo } from '@mnemoai/core';

// Auto-detect (uses OPENAI_API_KEY from env)
const mnemo = await createMnemo({ dbPath: './my-memory-db' });

// Preset
const mnemo = await createMnemo({ preset: 'ollama', dbPath: './db' });

// Full config
const mnemo = await createMnemo({
  embedding: {
    provider: 'openai-compatible',
    apiKey: process.env.OPENAI_API_KEY,
    model: 'text-embedding-3-small',
    dimensions: 1536,
  },
  dbPath: './my-memory-db',
});

Parameters

config.preset (optional)

Use a preset instead of manual embedding config. Available: "openai", "ollama", "voyage", "jina".

Or omit both preset and embedding for auto-detection (checks OPENAI_API_KEY → VOYAGE_API_KEY → JINA_API_KEY → error).

config.embedding (optional, overrides preset)

Field	Type	Default	Description
provider	"openai-compatible"	—	Provider type
apiKey	string	—	API key (use "ollama" for local)
baseURL	string?	OpenAI URL	API endpoint
model	string?	"text-embedding-3-small"	Model name
dimensions	number?	1024	Vector dimensions

config.dbPath (required)

Path to the local database directory. Will be created if it doesn't exist.

config.storageBackend (optional)

type StorageBackend = "lancedb" | "qdrant" | "chroma" | "pgvector"

Default: "lancedb" (embedded, zero-config).

config.decay (optional)

See Weibull Decay for details.

config.retrieval (optional)

See Retrieval Pipeline for details.

Errors

Error	Cause
mnemo: config is required	No config passed
mnemo: config.embedding is required	Missing embedding config
mnemo: config.embedding.apiKey is required	Missing API key
mnemo: config.dbPath is required	Missing database path