> ## Documentation Index
> Fetch the complete documentation index at: https://docs.cyborg.co/llms.txt
> Use this file to discover all available pages before exploring further.

# CyborgDB Embedded

CyborgDB Embedded provides native libraries for Python and C++, enabling you to integrate confidential vector search directly into your applications. This deployment model gives you maximum control, performance, and security by running entirely within your infrastructure.

## Why Choose CyborgDB Embedded?

**Maximum Security**\
Your data never leaves your environment. Vector embeddings remain encrypted on your hardware with keys under your complete control.

**Ultimate Performance**\
Direct integration eliminates network overhead and serialization costs. Take advantage of local CPU/GPU resources and custom optimizations.

**Full Control**\
Configure every aspect of the system - from index parameters to memory management. Perfect for custom requirements and specialized deployments.

**Any Environment**\
Deploy in air-gapped networks, edge devices, or custom infrastructure where external APIs aren't suitable.

***

## Available Libraries

<CardGroup cols={2}>
  <Card title="cyborgdb-core (Enterprise)" icon="star">
    *Full-featured production library*

    * Python + C++ bindings
    * All index types (IVFFlat, IVFPQ, IVF)
    * GPU acceleration with CUDA
    * Unlimited vectors and CPU threads
    * All backing stores (PostgreSQL, Redis, Memory)
    * Enterprise support
  </Card>

  <Card title="cyborgdb-lite (Evaluation)" icon="code">
    *Lightweight evaluation version*

    * Python bindings only
    * IVFFlat index type
    * Up to 1M vectors, 4 CPU threads
    * PostgreSQL and Memory backends
    * Perfect for testing and prototyping
    * Community support
  </Card>
</CardGroup>

<Note>
  **Free API Key Users**: If you're using a free API key, CyborgDB Service will automatically use [`cyborgdb-lite`](../../../embedded/guides/intro/about#available-libraries) under the hood, which includes:

  * Up to 1M vectors maximum
  * 4 CPU threads limit
  * PostgreSQL and Memory backing stores only
  * No GPU acceleration

  Upgrade to a [paid plan](https://www.cyborg.co/pricing) to unlock `cyborgdb-core` with unlimited vectors, all backing stores, GPU acceleration, and full performance capabilities.
</Note>

***

## Quick Start Path

<CardGroup cols={1}>
  <Card title="Embedded Quickstart" href="/versions/v0.11.x/embedded/guides/intro/quickstart" icon="rocket">
    *Get running in 5 minutes with Python or C++*

    Step-by-step guide covering both Python and C++ embedded library setup and usage.
  </Card>
</CardGroup>

## Architecture Overview

CyborgDB Embedded integrates directly into your application process:

```mermaid theme={null}
flowchart TB
  subgraph Your_Applications [Your Application]
    Library("CyborgDB Library (Python/C++)")
  end

  Library --> PostgreSQL("PostgreSQL (local or managed)")
  Library --> Redis("Redis (local or managed)")
  Library --> Memory("Memory (for development)")
```

**Key Benefits:**

* No external API dependencies
* Sub-millisecond query latency potential
* Complete data sovereignty
* Custom integration possibilities

***

## When to Use Embedded vs Service

<Tabs>
  <Tab title="Choose CyborgDB Embedded">
    * **Data must stay on-premises** - Regulatory or security requirements
    * **Ultra-low latency needed** - Sub-millisecond response requirements
    * **Custom integration required** - Unique workflow or system requirements
    * **Cost optimization critical** - High-volume usage with cost sensitivity
    * **Air-gapped deployment** - No external network access available
    * **Single application focus** - One primary application using vector search
  </Tab>

  <Tab title="Consider CyborgDB Service">
    * **Multi-language teams** - Need SDKs in multiple languages
    * **Microservice architecture** - Want to separate vector operations from application logic
    * **Operational simplicity** - Prefer service-based deployment patterns
    * **Independent scaling** - Need to scale vector operations separately
    * **Team collaboration** - Multiple teams need access to the same vector data
  </Tab>
</Tabs>

## Next Steps

<CardGroup cols={3}>
  <Card title="Start Building" href="/versions/v0.11.x/embedded/guides/intro/quickstart" icon="rocket">
    *Get hands-on with the quickstart guide*

    Follow our comprehensive guide covering both Python and C++ setup
  </Card>

  <Card title="Browse Capabilities" href="../encrypted-indexes/introduction" icon="code">
    *Explore detailed integration examples*

    Learn about encrypted index operations and data management
  </Card>

  <Card title="Framework Integration" href="../../../integrations/langchain/introduction" icon="plug">
    *LangChain and other integrations*

    Drop-in replacement for existing vector stores
  </Card>
</CardGroup>

Ready to build confidential AI applications with complete control? Start with our [Embedded Quickstart](/versions/v0.11.x/embedded/guides/intro/quickstart) guide!