MongoDB

The MongoDB plugin provides two capabilities: MongoPlugin registers Atlas Vector Search retrievers/indexers for RAG, and MongoSessionStore persists server-managed agent sessions.

Installation

<dependency>
    <groupId>com.google.genkit</groupId>
    <artifactId>genkit-plugin-mongodb</artifactId>
    <version>1.0.0-SNAPSHOT</version>
</dependency>

Requirements

• MongoDB 4.4+ (Atlas Vector Search requires MongoDB Atlas or the mongodb/mongodb-atlas-local Docker image)
• Java 21+

Vector search

MongoPlugin registers a retriever and indexer named mongodb/<collectionName> for each configured collection, backed by an Atlas Vector Search index and the $vectorSearch aggregation stage. A plain MongoDB server does not support $vectorSearch — use a MongoDB Atlas cluster or the mongodb/mongodb-atlas-local Docker image.

import com.google.genkit.plugins.mongodb.MongoPlugin;
import com.google.genkit.plugins.mongodb.MongoVectorStoreConfig;

Genkit genkit = Genkit.builder()
    .plugin(GoogleGenAIPlugin.create(apiKey))
    .plugin(
        MongoPlugin.builder()
            .connectionString("mongodb://localhost:27017/?directConnection=true")
            .addCollection(
                MongoVectorStoreConfig.builder()
                    .collectionName("films")
                    .embedderName("googleai/gemini-embedding-001")
                    .dimension(768)
                    .similarity(MongoVectorStoreConfig.Similarity.COSINE)
                    .createIndexIfNotExists(true) // create the Atlas Vector Search index on first use
                    .build())
            .build())
    .build();

// Index and retrieve
genkit.index("mongodb/films", documents);
List<Document> results = genkit.retrieve("mongodb/films", "a Christopher Nolan sci-fi film");

Tune per-collection settings with MongoVectorStoreConfig (database/collection names, embedder, index name, dimension, similarity — COSINE/EUCLIDEAN/DOT_PRODUCT, text/embedding field names, numCandidates, createIndexIfNotExists). See the mongo-vector sample.

Session store

Construct a MongoSessionStore from a com.mongodb.client.MongoClient and pass it to an agent’s .store(...) to persist server-managed sessions in MongoDB:

import com.google.genkit.plugins.mongodb.session.MongoSessionStore;
import com.google.genkit.plugins.mongodb.session.MongoSessionStoreOptions;
import com.mongodb.client.MongoClient;
import com.mongodb.client.MongoClients;

MongoClient client = MongoClients.create("mongodb://localhost:27017");

MongoSessionStore<Map<String, Object>> store =
    new MongoSessionStore<>(client); // uses database "genkit", collection "genkit_sessions"

Agent<Map<String, Object>> agent = genkit.beta().defineAgent(
    AgentConfig.<Map<String, Object>>builder()
        .name("myAgent")
        .system("You are a helpful assistant.")
        .store(store)
        .build());

All records live in a single collection whose _id is <prefix>::<recordId>; the database and collection are created automatically on first write. It uses the same sharded checkpoint + RFC-6902 diff + pointer layout as the Firestore, DynamoDB, Cosmos DB, and PostgreSQL backends, and supports onSnapshotStateChange (via polling), so chat.abort() works.

See Session Stores for options.

Sample

See the agents-mongo-session sample and the mongo-vector sample.