Building on a trusted data foundation to support scale, context, and AI adoption
As we head into 2026, it’s a good moment to reflect on how Arango evolved over the past year. In 2025, our focus was twofold:
- strengthening ArangoDB as a high-performance, multi-model database, and
- expanding the platform around it to support emerging, AI-driven use cases.
Everything we shipped built on the same proven foundation—unified graph, vector, document, key-value, and search—while adding better tooling, cloud-native operations, and optional AI capabilities that work directly with your data. What started as foundational improvements early in the year became a more unified platform by year’s end, giving teams the flexibility to adopt new capabilities at their own pace.
Below is a look at the key milestones we delivered in 2025—and what they enable you to build today.
TL;DR
In 2025, Arango strengthened its core ArangoDB multi-model database while evolving it into a more unified, cloud-native platform. Early in the year, the focus was on performance, developer productivity, and Kubernetes-native operations.
In the second half, Arango introduced optional, graph-powered AI capabilities—culminating in the October pre-release of the Arango AI Data Platform, which combines ArangoDB, unified tooling, and AI features like VectorRAG, GraphRAG, HybridRAG, graph analytics, and bring-your-own LLMs. The result is a flexible platform that supports both traditional production workloads and emerging AI-driven use cases, letting teams adopt new capabilities at their own pace as Arango heads into 2026.
2025 Highlights at a Glance
ArangoDB
Enhanced querying, graph exploration, and developer productivity through a redesigned web experience and expanded visualization capabilities.
Platform Experience (Built on ArangoDB)
Introduced a unified, Kubernetes-native platform layer for deploying, operating, and scaling ArangoDB with improved reliability, observability, and control.
Arango AI Data Platform
Launched in October as an extension of the platform, adding optional, graph-powered AI capabilities—including VectorRAG, GraphRAG, and HybridRAG, AIOps & MLOps, and natural language querying—to help teams work with contextual data.
Our 2025 Product Journey
1H 2025
Strengthening the ArangoDB Foundation
In the first half of the year, we focused on strengthening the core ArangoDB experience and the platform capabilities required to run reliably at scale. This work laid the groundwork for a more unified, cloud-native platform—while continuing to support teams using ArangoDB as a high-performance multi-model database for production applications.
Key areas of investment included:
- ArangoDB’s multimodel capabilities (graph, vector, document, key-value, and search)
- Platform services such as high availability, monitoring, APIs, and connectors
- Kubernetes-native deployment, scaling, and lifecycle management
2H 2025
Expanding Capabilities for Contextual Data and AI
As teams began exploring AI-driven use cases on top of their existing ArangoDB data, we introduced an initial set of optional, graph-powered capabilities—later incorporated into the Arango AI Data Platform.
These included:
- GraphRAG for turning unstructured data into contextual knowledge graphs
- Graph Analytics (PageRank, Connected Components, and more)
- GraphML for machine learning and embedding generation
- Jupyter Notebooks with pre-configured ArangoDB and data science tooling
- MLflow integration for experiment tracking and model management
- Triton Inference Server for secure, on-premises LLM deployment
Q4 2025
Pre-Release of the Arango AI Data Platform
In October, at the Nvidia GTC AI Conference, we launched the Arango AI Data Platform—bringing together ArangoDB, a unified web experience, Kubernetes-native orchestration, and optional AI capabilities into a single, integrated platform.
The platform builds directly on ArangoDB and includes:
- ArangoDB as the multi-model data foundation
- A unified web interface for database, platform, and AI workflows
- Kubernetes-native operations powered by the ArangoDB Operator
- Built-in access to the AI Suite for graph-powered AI use cases
Explore the Arango AI Data Platform →
Applied Solutions and Nvidia Ecosystem Momentum
To help teams move from experimentation to production, we released solution blueprints and ecosystem integrations, including:
- An AI Blueprint for Video Search and Summarization using GraphRAG and ArangoDB
- Accelerated, production-ready graph analytics for NetworkX users
These releases demonstrated how ArangoDB and graph-powered analytics can be applied to practical, industry-specific use cases.
- AI Blueprint for Video Search and Summarization (VSS)
A reference architecture for deploying video analytics AI agents using GraphRAG and ArangoDB across industries.- Check out Nvidia Docs: VSS Blueprint Architecture
- Read the blog post: Video Search & Summarization Blueprint with GraphRAG & ArangoDB
- Watch the Video: Search & Summarization Blueprint demo
- Accelerated, Production-Ready Graph Analytics for NetworkX Users
Delivered high-performance graph analytics for NetworkX workloads, enabling teams to scale graph algorithms from notebooks to production.- Read the blog post: Accelerated, Production-Ready Graph Analytics for NetworkX Users
Platform Experience Enhancements (Built on ArangoDB)
Query Editor Improvements
A modernized AQL Query Editor to improve productivity, collaboration, and performance analysis.
Graph Visualizer Enhancements
Expanded visualization, customization, and performance to support deeper analysis at scale—including large graphs with millions of nodes and edges.
AI Capabilities (Platform-Included, Optional)
GraphRAG Enhancements
- Instant and Deep Search modes
- Incremental updates to existing knowledge graphs
- Improved workflows for working with evolving data
Unified LLM Configuration
- Support for OpenAI-compatible APIs (OpenAI, OpenRouter, Gemini, Anthropic) and company-built LLMs self-hosted models
hosted and self-managed LLMs - Flexible mixing of providers for chat and embeddings
AQLizer — Natural Language to AQL
- Query ArangoDB using natural language
- Generate AQL directly in the Query Editor
- LLM-powered insights built directly into developer workflows
Looking Ahead to 2026
2025 was a meaningful year for Arango—strengthening ArangoDB as a trusted multi-model data foundation while expanding what teams can build on top of it as their needs evolve.
Whether you rely on ArangoDB today to power production applications or are beginning to explore AI-driven use cases, everything we shipped was designed to build on the same proven core.
As we move into 2026, our focus remains the same: helping teams work with contextual data more effectively, operate with confidence at scale, and adopt new capabilities at their own pace. We’re grateful for the customers, partners, and community members who helped shape what we delivered, and we’re excited to continue building together in the year ahead.
Explore what’s now available in 2025 Product Updates →
