ArcadeDB

ArcadeDB is an open-source multi-model database for applications that need to work with graph, document, key-value, full-text search, vector, and time-series data in one engine. It supports ACID transactions in single-node and distributed deployments. Developers can query it with SQL, OpenCypher 25, Gremlin, GraphQL, MongoDB query language, or a native Java API. It can run embedded in Java applications or as a standalone server, with installation options including Docker, a GitHub binary, Maven, and Kubernetes. ArcadeDB Server supports HTTP/JSON and compatibility with Postgres wire protocol, Redis, and MongoDB protocols. Built-in importers are available for Neo4j and OrientDB, and the maker says Raft consensus maintains consistency across multiple servers. The database is free for any usage under the Apache 2.0 license, with no usage limits or feature gating. Paid support subscriptions start at $299 a month per server, with response times and coverage varying by plan. A separate Studio AI Assistant license is $19.99 a month.

ArcadeDB brings multiple data models and query interfaces into one database, with free usage under Apache 2.0. Teams needing guaranteed support response targets can choose a paid plan, while the database itself has no listed usage limits or feature gating.

ArcadeDB may suit developers building knowledge graphs, GraphRAG, fraud detection, recommendation engines, real-time analytics, customer 360, supply-chain analytics, or AI/ML feature stores. Its multiple query languages and deployment options also fit teams that need to work with several data models or deployment styles.

Overview

ArcadeDB is an open-source database that combines graph, document, key-value, full-text, vector, and time-series data in one engine. It suits teams building graph-heavy, analytics, or AI-related systems that want control over deployment and database features. Its main appeal is a full-featured Apache 2.0 edition; the trade-off is that support with response-time commitments costs extra.

That breadth can reduce the need to operate separate database types, but it is not an automatic advantage for a workload that needs only one narrow model. ArcadeDB is led by Arcade Data Ltd, a private company headquartered in London, United Kingdom. For related categories, see Multimodel Databases, Graph Databases, and Database Software.

Key features

• Multiple models, plus graph and vector tools: Graph algorithms and vector search sit alongside document, key-value, full-text, and time-series data. That combination suits knowledge graphs, GraphRAG, fraud detection, recommendations, and AI/ML feature stores; projects with a single, straightforward data need may not benefit from the extra breadth.
• Several query interfaces: SQL, OpenCypher 25, Gremlin, GraphQL, MongoDB query language, and a native Java API give teams choices across application styles. This can ease adoption in mixed environments, though teams still need to decide which interface best fits each workload.
• Flexible deployment and compatibility: ArcadeDB can run embedded in Java applications or as a standalone server, with installation through Docker, a GitHub binary, Maven, or Kubernetes. Its server supports HTTP/JSON and compatibility with the Postgres wire protocol, Redis, and MongoDB protocols, offering integration routes without limiting deployment to one style.
• Transactions and distributed consistency: ArcadeDB states that it provides full ACID transactions for single-node and distributed deployments and uses Raft to maintain consistency across multiple servers. A persistent write-ahead journal and automatic rollback of uncommitted transactions after a crash address recovery concerns, but these claims do not remove the need to plan and operate a deployment carefully.
• Migration paths: Built-in importers for Neo4j and OrientDB may help teams move existing data into ArcadeDB rather than starting from scratch.

Pricing

The database itself is free for any usage under Apache 2.0, with no usage limits or feature gating. That makes the Community plan unusually straightforward for production use: the price is free, billed forever, and it includes the full ArcadeDB feature set, community forums, GitHub Issues, public documentation, and tutorials. The cost is that support is community-based rather than backed by a paid response-time SLA.

• Silver — $299 a month: Billed monthly per server. It adds email support, a priority issue queue, backport patches, a 4h S1 response during business hours, and one Studio AI Assistant license. It fits teams that need a defined support route without paying for round-the-clock coverage.
• Gold — $599 a month: Billed monthly per server. It raises the S1 response target to 1h, includes 8am–8pm CET coverage and 24/7 coverage, guaranteed backport patches, escalation management, and one Studio AI Assistant license. This is more suitable when faster response and escalation are worth the per-server premium.
• Platinum — $1,499 a month: Billed monthly per server. It includes a 30min S1 response, 24/7/365 critical coverage, a dedicated technical account manager, root-cause analysis for S1/S2, an annual architecture and performance audit, and one Studio AI Assistant license. It is aimed at deployments where intensive, continuous support justifies the highest recurring cost.
• Studio AI Assistant — $19.99 a month: Billed monthly per license. Natural-language query assistance, schema exploration, query optimization suggestions, and context-aware documentation are useful for teams seeking that guidance separately; this license does not replace a support plan's response commitments.

Support plans are billed per server, while the assistant is billed per license. No seat or server caps are given beyond those billing units, so costs will rise with the number of covered servers or assistant licenses.

Platforms

ArcadeDB supports API, Linux, macOS, self-hosted, web, and Windows environments. The latest release page lists version 26.9.1 with Linux, macOS, and Windows downloads. Embedded Java deployment and standalone server operation make it suitable for teams that want to choose between application-level integration and a separate service.

Who it's for

ArcadeDB is a strong fit for teams consolidating graph, vector, document, and other data models, particularly for use cases such as customer 360, supply-chain analytics, real-time analytics, and recommendation engines. The free, unrestricted core is also attractive to developers who can rely on community support.

Pros and cons

• Pro: Apache 2.0 licensing and no usage limits or feature gating keep the full database available without a paid core tier.
• Pro: Multiple data models, query languages, and deployment modes give teams room to fit different workloads and existing systems.
• Pro: ACID transactions, Raft-based consistency, and built-in importers for Neo4j and OrientDB address important operational and migration needs.
• Con: Guaranteed support response targets require per-server subscriptions starting at $299 a month, making support costs scale with deployment size.
• Con: The broad model and interface set may be unnecessary overhead for applications that need only one focused database capability.

Alternatives

Amazon DynamoDB is worth considering when a managed cloud database with a free tier is a better fit: that tier includes 25 WCUs, 25 RCUs, 25 GB storage, and 25 replicated WCUs per region and payer account, unlike ArcadeDB's unrestricted self-managed database core.

CrateDB is an alternative for teams seeking a freemium database with a defined shared free allocation of 2 vCPUs, 2 GiB RAM, and 8 GiB storage.

MarkLogic may suit noncommercial development where a developer plan capped at 1TB of data is appropriate.

RavenDB offers a developer plan specified for 9 cores, 36 GB RAM, and 3 nodes, which may better match teams comparing against a bounded developer allocation.

Aerospike Database has a Community Edition specified for 8 nodes, 2.5 TB of data, and 2 namespaces, a clearer choice when those limits suit the deployment.

AllegroGraph is an option for graph workloads that fit its free edition's maximum of 5,000,000 triples.

Azure Cosmos DB for NoSQL may suit teams seeking a free tier with 1,000 RU/s, 25 GB storage, and one free-tier account per Azure subscription.

InterSystems IRIS is another alternative with a free Community Edition.

Verdict

Choose ArcadeDB if your application benefits from several data models in one engine and you want the freedom to use the full database without a license-based feature ceiling. Its strongest case is breadth paired with an unrestricted Apache 2.0 core. Look elsewhere if a specialized database better matches a narrow workload, or if your team wants a different deployment model; choose a paid ArcadeDB support plan when guaranteed response times matter.