A lean, high-performance distributed database that uses object storage (MinIO, S3, Azure Blob Storage) as its storage layer. Built for massive scale at minimal cost.
Azure blob storage support is upcoming!
- Cost-efficient at scale: Leverage object storage economics to store petabytes at a fraction of traditional database costs
- Inherited durability: Built-in fault tolerance, redundancy, and geographic distribution from object storage providers
- Zero infrastructure overhead: No sharding, replication, or cluster management required—object storage handles it all
- Global distribution: Deploy instances worldwide with NATS providing seamless, location-transparent connectivity
Built from scratch in Go with minimal resource footprint. Uses NATS as the communication layer for effortless global deployment—hundreds of instances, zero load balancers, no DNS complexity.
Nimbus excels for workloads matching these patterns:
-
Point lookups by ID: Fast, direct retrieval of individual objects. i.e. get object with some id. e.g. getting medical results by visit id. or getting full information of a medical diagnosis test by test id etc.
- People often underestimate this usecase. It is quite versatile and many workloads can be fit into this.
-
Temporal streaming: Stream objects from collections in insertion timestamp order. e.g. get all activities performed by a user.
Ideal use cases: Event sourcing, audit trails, time-series logs, document stores, and any workload where data is append-heavy with lookups by a single Id.
If you have a workload that can fit into these criteria, you can benefit by using Nimbus to achieve massive scale at extremely low cost.
See API Doc
- For running the project locally, see Local Setup
- If you are interested in contributing, coding standards and style guidelines, see the Contribution Guide.
- For Nimbus architecture, please see Architecture & Dev Docs
🏗️ Architected by Gaurav Kalele,
💫 Vibe coded with Cursor
👮 Thoroughly Reviewed by: Gaurav (Human), Cursor (self review), Copilot (PR Review)
© Connection Loops