Managed Mode

In managed mode, Embedd.to handles vector storage for you using its built-in Qdrant vector database.

How It Works

1. You create a connection to your source database
2. You register an embedding provider (OpenAI, Gemini)
3. You create a vector table — Embedd.to creates a Qdrant collection automatically
4. On backfill, source data is read, embedded via your embedding provider, and stored in Qdrant
5. Sync keeps vectors up to date as source data changes

When to Use Managed Mode

• Getting started — Fastest way to add semantic search
• Multi-provider search — Query vectors from multiple database providers through one API
• No infrastructure changes — No need to install pgvector or manage vector indexes

Requirements

• A source database connection (Snowflake or PostgreSQL)
• An embedding provider with a valid API key

Architecture

Source DB → Embedd.to → Embedding Provider (OpenAI/Gemini)
                ↓
           Qdrant (managed by Embedd.to)
                ↓
           Query API

Filter Support

Managed mode uses Qdrant's native filtering with full support for all filter operators ($eq, $ne, $gt, $gte, $lt, $lte, $in, $nin, $exists). Filters are translated to Qdrant's FieldCondition format automatically.

Limitations

• Vectors are stored in Embedd.to's infrastructure, not yours
• Query latency depends on Qdrant instance proximity to your application