A custom skill designed to provide expert guidance and code patterns for building applications with Qdrant Edge — the embedded, offline-capable vector search engine for edge devices.
Note & Disclaimer:
- This is not an official Qdrant skill.
- It was prepared based on the Qdrant Edge beta version. Please be aware that APIs and functionalities may undergo changes, so it's always good to check the official docs.
- This skill was generated using Claude Sonnet 4.6 Extended Thinking with the following prompt:
I want to create a “skill” to be used in coding agents such as Claude Code, Gemini CLI, and Antigravity for Qdrant Edge. The relevant sources are: https://qdrant.tech/documentation/edge/ https://qdrant.tech/documentation/edge/edge-quickstart/ https://qdrant.tech/documentation/edge/edge-fastembed-embeddings/ https://qdrant.tech/documentation/edge/edge-data-synchronization-patterns/ https://qdrant.tech/documentation/edge/edge-synchronization-guide/ https://github.com/qdrant/qdrant/tree/master/lib/edge/python/examples Create the necessary “skill” accordingly.
This skill equips AI assistants with deep knowledge about qdrant-edge-py, enabling them to help developers implement efficient, on-device vector search without requiring network connectivity. It covers local storage operations, on-device embedding generation with fastembed, and robust synchronization patterns between edge clients and centralized servers.
This skill contains comprehensive documentation and code patterns for:
- Core Workflow: Initialization, configuring local
EdgeShardinstances, and basic CRUD operations (upsert, query, retrieve, scroll). - On-Device Embeddings: Generating embeddings entirely locally (without internet) using
fastembedand integrating them directly into the Qdrant Edge instance. - Data Synchronization: Server-to-Edge synchronization using full and partial snapshots, and Edge-to-Server asynchronous dual-write patterns.
- Best Practices: Avoiding common pitfalls, ensuring proper resource closure, offline-first design, and managing edge constraints.
SKILL.md: The core metadata and primary guide that triggers the skill, containing the essential configuration and code snippets for Qdrant Edge.references/fastembed.md: Detailed guidance on usingfastembedto generate text and image embeddings directly on edge devices.references/synchronization.md: Complete synchronization patterns, demonstrating how to keep edge data in sync with a remote Qdrant server.
The skill is automatically invoked when discussing:
Qdrant Edgeqdrant-edge-pyEdgeShardon-device vector searchoffline vector searchembedded vector database- Synchronizing vector data between an edge device and a remote server.