Here is wishing our readers an exciting new year as we come out of discussing whether we need agents into how we productionize them with a healthy dose of context and multimodality, in 2026! As we reflect on all we have achieved at ApertureData this year, we are extremely proud of our team and very thankful to our current as well as potential customers, investors, and families in supporting us as the AI landscape  continues to evolve rapidly (read our state of the union report from data & AI leaders)! 

TL;DR

• ApertureDB platform now has AI workflows and plugins on top of your favorite vector-graph-multimodal database to make AI/ML application development and production deployment easy.
• ApertureDB is 2-35X faster than alternatives (and we have a detailed performance white paper breaking it down).
• ApertureDB is available in more environments to be deployed wherever you are, with an SDK that simplifies all your ETL challenges, all with the right security & compliance guarantees at scale.
• We have always believed in the power of graphs for context and building a knowledge base. Looks like the research and product community finally agrees with us wholeheartedly. We are happy to support building knowledge graphs on ApertureDB efficiently and at scale.
• Agentic memory support at scale for enterprise AI agents is the next frontier for us

Some highlights from 2025

What ApertureDB can do for you: ApertureDB is the only database that seamlessly combines the capabilities of a memory-optimized graph database with scalable vector database and multimodal dataset management. You only have to work with one database via its unified API. In addition to VPC deployments you can use ApertureDB with its free community edition, GCP marketplace, and self-serve on GCP as well as AWS (see all options in documentation)!

AI workflows and plug-ins: We introduced AI workflows to simplify tasks like embeddings generation, RAG implementations, even labeling, and more to make it very easy to create entire applications beyond just managing data. In fact, we now have plugins like MCP server, SQL server, and Jupyter notebook to serve entire AI and analytics teams. You can build an entire RAG application, with just a few clicks!

‍Client improvements: Let’s face it, it was not easy to connect and use ApertureDB in 2024! We learnt that the hard way in a hackathon earlier in 2025. Now you have a very simple token based access mechanism. Add it to the environment, and it works everywhere. Couple that with the workflows, and multimodal AI just becomes so straightforward. Our SDK got some other improvements like a NodeJS interface, new Python object API for videos and embeddings, as well as much better loader support to run large scale workloads. ApertureDB UI now supports PDFs, text vector search, and semantic search across various modalities. Our LlamaIndex integration now includes graph search too! And Label Studio can be used easily as an AI workflow.

Scale & Performance: Not only were we able to scale ApertureDB to 1M queries per second, measure some of the best search performance in the market (13000+ embeddings per second, sub 10 msec vector search latency, sub 15msec lookup in a Billion scale graph), we can also scale to where other popular vector-graph databases fall short, as captured in this performance report!

Growing community: Did you get to join our lunch & learn sessions packed with practical examples and AI updates from our partners? We have been loving the community growth and contributions from our valuable community members on creating Graph RAG, Agentic RAG, AI agents, secure RAG, and other such interesting examples. We were overwhelmed with feedback from the couple of hackathons we sponsored this year, a number of talks, and panels we got invited to, resulting in insightful content around data and data tools relevant in the AI world.

What to expect from ApertureData in 2026?

We will continue the efforts we started in 2025 and work with you every step of the way to solve the evergrowing collection of interesting text and multimodal data challenges!

A few new efforts that we are very excited about:

• Agentic memory development, all in open source (follow our blog series on agent memory)
• We have lined up more exciting partnerships around agentic workflows, large scale storage, ETL, and some very interesting new content (see also our community page)
• New datasets, use case examples, and updated performance report (have you checked out our movie dataset workflow yet?)
• Petabyte scale data cloud deployments

Do follow our newsletter and join us for various events in 2026! If you would like to evaluate ApertureDB on your own, please go ahead and try our Quick Start guide.