Skip to content

pgvector Storage

For common Agent integration, extraction modes, and tool configuration, see Usage and Configuration.

Use case: Production, vector similarity search with PostgreSQL + pgvector

Set PGVECTOR_DSN through environment or secret management. A production DSN should validate the server certificate and use a host name covered by it:

1
postgres://<user>:<password>@db.example.com:5432/dbname?sslmode=verify-full&sslrootcert=<trusted-ca-path>

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
import (
    "os"

    openaiembedder "trpc.group/trpc-go/trpc-agent-go/knowledge/embedder/openai"
    memorypgvector "trpc.group/trpc-go/trpc-agent-go/memory/pgvector"
)

embedder := openaiembedder.New(openaiembedder.WithModel("text-embedding-3-small"))

pgvectorService, err := memorypgvector.NewService(
    memorypgvector.WithPGVectorClientDSN(os.Getenv("PGVECTOR_DSN")),
    memorypgvector.WithEmbedder(embedder),
    memorypgvector.WithSoftDelete(true),
)
if err != nil {
    panic(err)
}

Configuration options:

  • WithPGVectorClientDSN(dsn): Recommended connection form; has the highest priority
  • WithHost/WithPort/WithUser/WithPassword/WithDatabase: Alternative field-based connection parameters
  • WithSSLMode(mode): SSL mode (default "disable")
  • WithPostgresInstance(name): Use pre-registered PostgreSQL instance
  • WithEmbedder(embedder): Text embedder for vector generation (required)
  • WithSoftDelete(enabled): Enable soft delete (default false)
  • WithTableName(name): Custom table name (default "memories")
  • WithSchema(schema): Specify database schema (default is public)
  • WithIndexDimension(dim): Vector dimension (default 1536)
  • WithMaxResults(limit): Max search results (default 10)
  • WithMemoryLimit(limit): Memory limit per user
  • WithCustomTool(toolName, creator): Register custom tool
  • WithToolEnabled(toolName, enabled): Enable/disable tool
  • WithExtraOptions(...options): Extra options passed to PostgreSQL client
  • WithSkipDBInit(skip): Skip table initialization (for users without DDL permissions)
  • WithHNSWIndexParams(params): HNSW index parameters for vector search

Note: A DSN takes priority over the field-based connection parameters. Both direct forms take priority over WithPostgresInstance. The default SSL mode is disable; use it only for trusted local development, not production. The pgvector extension must be installed in PostgreSQL.

Default initialized schema:

The service enables the vector extension and initializes public.memories with vector, episodic, and full-text fields. WithTableName, WithSchema, and WithIndexDimension replace memories, public, and 1536. It also creates indexes for app/user, update time, deletion time, event time, kind, participants (GIN), embedding (HNSW), and search_vector (GIN), plus a trigger that maintains search_vector from memory_content.

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
CREATE EXTENSION IF NOT EXISTS vector;

CREATE TABLE public.memories (
    memory_id TEXT PRIMARY KEY,
    app_name TEXT NOT NULL,
    user_id TEXT NOT NULL,
    memory_content TEXT NOT NULL,
    topics TEXT[],
    embedding vector(1536),
    memory_kind TEXT NOT NULL DEFAULT 'fact',
    event_time TIMESTAMP NULL,
    participants TEXT[],
    location TEXT NULL,
    created_at TIMESTAMP NOT NULL DEFAULT CURRENT_TIMESTAMP,
    updated_at TIMESTAMP NOT NULL DEFAULT CURRENT_TIMESTAMP,
    deleted_at TIMESTAMP NULL DEFAULT NULL,
    search_vector tsvector
);

WithSkipDBInit(true) skips the extension, table, indexes, trigger function, trigger, and full-text backfill. Provision all of them before starting the service; use memory/pgvector/init.go as the authoritative DDL, including custom HNSW parameters.

Resource cleanup: Call Close() method to release database connection:

1
defer pgvectorService.Close()