JayDB

Overview

JayDB combines per-user sign-in with a JSON document store, giving static and AI-built apps a way to keep user, session, or tenant data without an app-owned backend. It is a focused choice for apps that can address documents by path; workloads built around querying relationships or updating several records together need a different database.

Key features

Path-based documents and HTTP

Documents live at slash-separated paths, and the first write creates the namespace automatically. The data plane uses standard REST HTTP for GET, PUT, DELETE, and prefix listing, with JSON, MsgPack, and Raw document formats. This gives clients a compact set of ways to read and change known documents, but JayDB has no secondary indexes, joins, foreign-key constraints, full-text search, or aggregation. It also has no vector search or transactions, so it is a poor fit for flexible discovery, relational data, or workflows requiring atomic changes across documents. Documents can be up to 5 MB.

Sign-in and SDK

Each tenant gets an OIDC issuer, and browser apps can use Authorization Code with PKCE without putting a client secret in the app. Supported identity integrations include Google, GitHub, Microsoft, generic OIDC, and custom OAuth2 providers. The official zero-dependency @jaydb/cloud SDK handles OIDC and PKCE authentication, token refresh, typed document access, and compare-and-swap operations. That combination is useful when an app needs user-scoped storage and wants to avoid building its own sign-in and data service.

Concurrency and hosting

ETag-based compare-and-swap can reject a write with HTTP 412 if a document changed after it was read, helping clients avoid silently overwriting one another. It does not provide multi-document atomicity, and sustained concurrent writes to a single hot key can cause retries to pile up. The managed cloud option uses AWS S3 or compatible storage for durability. The MIT-licensed Go engine can instead be embedded or run as a standalone server with S3, Cloudflare R2, MinIO, or local disk storage, giving teams a self-hosting route as well as managed hosting.

Pricing

JayDB is freemium, with no free trial. The Free plan costs nothing, billed Free every month, forever. It includes 1 GB stored, 100,000 reads per month, and 10,000 writes per month; usage hard-stops at those allowances. That is a useful way to start a small app, but the hard stops make it unsuitable for workloads that cannot tolerate an exhausted quota.

For paid usage, storage costs $0.05 per GB-month, reads cost $0.20 per million, and writes cost $0.50 per million. Deletes are free, and egress has no per-GB line item under fair use. The $5 prepaid pack costs $5, paid once and prepays $6.50 of metered usage, a 1.3× discount. The $10 prepaid pack costs $10, paid once and prepays $15 of metered usage, a 1.5× discount. These packs suit users who want a discounted balance for the same metered usage rather than a different feature tier.

Platforms

JayDB supports Android, iOS, Linux, macOS, Windows, web, and API use, as well as self-hosted deployment. Its managed-cloud and self-hosted options let teams choose between outsourcing operation and running the Go engine themselves.

Who it's for

JayDB is best suited to AI-built apps, multi-user or worker-shared documents, and state tied to a user, session, or tenant. It makes sense when data can be organized into individual documents and accessed through known paths. Choose another database when search, joins, indexes, aggregation, vector search, or atomic multi-document updates are core requirements.

Pros and cons

• Per-tenant OIDC and PKCE: browser apps can use sign-in without shipping a client secret, with several identity-provider options.
• Small, standard HTTP surface: REST operations and automatic namespace creation suit apps that need direct document access without a custom backend.
• Managed or self-hosted: the MIT-licensed Go engine can use S3-compatible storage or local disk.
• Hard limits on the free plan: reaching a monthly allowance stops usage rather than simply moving it onto metered billing.
• Narrow data model: no indexes, search, joins, aggregation, or cross-document transactions rules out many query-heavy and consistency-sensitive applications.
• Hot-key contention: sustained concurrent writes to one document can lead to accumulating retries.

Alternatives

Document Databases is a useful starting point for comparing other document-store options.

• PouchDB is a free option with Android, iOS, Linux, macOS, self-hosted, web, and Windows platforms.
• RavenDB is a freemium alternative with Developer and Community plans and support for API, Linux, macOS, self-hosted, web, and Windows.
• Apache CouchDB is free, open-source software for self-hosting on servers or VMs, as a single node or cluster.
• eXist-db is a free open-source option for academic, non-commercial, and commercial applications.
• Firestore is a freemium alternative with a free trial and Android, API, iOS, and web support.
• MarkLogic is a freemium alternative with a free trial and Developer and Enterprise plans.
• Apache AsterixDB is free open-source software for a self-managed cluster.
• Azure DocumentDB is a freemium option with a Free Tier that includes 32 GB of storage and one free database server per subscription.

Verdict

Choose JayDB for an app that needs per-user sign-in and straightforward JSON documents, with the choice of managed hosting or self-hosting. Its focused REST model and tenant-specific identity are the main reasons to pick it; the lack of query features and cross-document transactions is the reason to look elsewhere when data relationships or coordinated updates matter.