Glider

Glider is a single-node vector database that uses S3-compatible object storage as its durable source of truth.

Glider is a single-node vector database designed to store and search vector data using S3-compatible object storage as the durable source of truth. Every acknowledged write is made durable in S3 before the server responds, while local RAM and SSD caches speed up read operations. It serves nearest-neighbor searches and metadata filters over an HTTP/JSON API, and provides a Python client along with an MCP memory server for AI agents. The database handles multiple collections with distinct dimensions and metrics, supports server-side embeddings via local models or OpenAI, and includes tools for backup and restore. It is intended for developers and applications requiring reliable vector search backed by object storage without complex multi-node orchestration.

Glider at a glance

Best for: Software engineer
Runs on: API, CLI, self-hosted, open source
Pricing: Free
Released: October 2026
Category: AI coding tool
Website: github.com/omerfeyzioglu/glider

Key features

Durable storage on S3 object storage
Crash-safe server takeover and log replay
Retry-safe writes with request IDs
Multiple collections with custom dimensions
Metadata filters for search queries
Background clustered ANN search
Local SSD cache for faster warm queries
HTTP/JSON API and Python client

Use it for

Storing and querying vector embeddings for AI applications
Providing durable memory for AI agents via MCP
Running nearest-neighbor search with metadata filters
Managing multiple vector collections on a single node

Alternatives to Glider

HelixDB — Combines graph, vector, and full-text search in one engine.
Pg Vectorsearch — Provides faster querying and building than standard indexes.
zvec — It is a pure local, serverless vector database providing fast hybrid search.
Weaviate — Stores data objects and vectors for semantic and hybrid search capabilities.
LEANN — Efficiently index millions of documents locally with 97% storage savings.
seekdb — Unifies relational, vector, and text data for production-grade AI agents.
CubeSandbox — Provides sub-60ms cold starts and hardware-level microVM kernel isolation.
pgEdge — It combines AI tooling with standard Postgres and flexible deployment options.
Open Glean — Provides agents with structured state and relationship-based context rather than just similarity.
LangChain4j — Integrates LLMs and vector stores directly into Java applications.
Daytona — It provisions sandboxes in under 90 milliseconds with full security isolation.
Codebase Memory MCP — High-speed indexing and persistent knowledge graphs for AI coding agents.

All Glider alternatives