Qdrant Edge

Edge AI Infrastructure

Run Vector Search Inside Embedded and Edge AI Systems

Qdrant Edge is a lightweight, in-process vector search engine designed for embedded devices, autonomous systems, and mobile agents. It enables on-device retrieval with minimal memory footprint, no background services, and optional synchronization with Qdrant Cloud.

Read Quick Start Docs Contact Us

Real-time vector retrieval for Edge AI in resource-constrained environments

Native Vector Search for Embedded & Edge AI

Run in-memory, disk-backed, and hybrid vector search on the edge. Deploy on mobile devices, IoT gateways, industrial PCs, drones, and more.

Optimized for Low Memory & Low Compute Devices

Optimized for resource-constrained environments with a small memory footprint and efficient CPU/GPU utilization to ensure smooth performance on edge devices.

Local-first, Cloud-Connected When Needed

Perform vector search locally with fallback to cloud for more complex queries or when more compute is needed to train your AI model.

Hybrid & Multi-modal Search On Device

Support for various data types, including text, images, audio, and more. Combine multiple modalities for more accurate and context-aware results.

Multitenancy Built for Edge Scale

Designed to manage multiple tenants, users, or applications on a single edge device. Isolate data and control access for secure and scalable deployments.

Purpose-Built for On-Device AI Workloads

Robotics & Autonomy

Run real-time vector search to enable robots to make faster, more informed decisions for object recognition, navigation, and more.

Offline Voice Assistants

Provide fast, accurate, and private voice search for devices without an internet connection, such as smart speakers, wearables, and more.

Smart Retail & Kiosks

Personalize in-store experiences, provide product recommendations, and power intelligent kiosks for enhanced customer engagement.

Industrial IoT

Perform anomaly detection, predictive maintenance, and real-time insights on sensor data directly at the edge for industrial applications.

Demo: Offline Visual Memory for Smart Glasses

This GitHub demo showcases a proof-of-concept for smart glasses that can remember what they see and help you find objects, like your keys, even when fully offline. It runs Qdrant Edge directly on the device, using a vision-language model to convert video frames into vectors for fast, local search while skipping redundant frames to stay efficient.

See how vector search can bring memory-like capabilities to resource-constrained hardware at the edge.

View in GitHub

Submit Your Interest or Project

Fill out the form to stay updated about Qdrant Edge news, or let us know about your project.

In filling out the Qdrant Edge beta form, I can confirm that:

We are building or deploying AI systems on embedded or edge devices (e.g. robots, mobile hardware, IoT, or offline agents)
We require local vector search as part of our product or infrastructure
We are able to test and provide feedback within the next 60 days

First Name*

Last Name*

Business Email*

Company name*

FAQs

Who is Qdrant Edge for?

Teams building AI systems that need fast, local vector search on embedded or resource-constrained devices, such as robots, mobile apps, or IoT hardware.

Is this available to all Qdrant users?

Yes. Read the Quick Start guide, and view the demo on GitHub.

How do I get access?

If you're building edge-native or embedded AI systems, apply to join the beta. Or, read the Quick Start guide, and view the demo on GitHub

Apply to Join the Beta

Apply Now