Listening to Bryan Adams: Our TypeDB Summer ’26 Roadmap

Listening to Bryan Adams’ Summer of ’69 got me feeling excited for the upcoming summer. Given it’s a quiet week in the office, with a bunch of people on holiday ☀️ (or sick 😥), it feels like a good time to look forward to our summer of ’26. I’m incredibly excited about what we’ve got planned, and I wanted to share a glimpse of it with our enthusiastic and dedicated community.

I want to start by thanking all of you for the wonderful projects that have been kicking off lately. From a Pydantic-based ORM (type-bridge) to AI focused projects like context management (AutoK) and knowledge notebooks (Skillful Alhazen), to the community Julia and Swift drivers currently in the works. From archival research to so much more – your projects and work are pure motivation and joy. You’re one of the main reasons I feel so enthused!

A big thank you also goes out to our commercial partners, who are doing exciting work across a multitude of fields, from fintech to government infrastructure management. Your work and input are critical to our success.

The Year So Far

We’ve been wrapping up a series of long-running projects and paying down technical debt. Now, we’re in a fantastic place to push forward with the next phase of features, performance, and ecosystem development that will make TypeDB even more enjoyable to use.

Here are a few pieces of the team’s work I want to highlight.

Infrastructure & Bazel Upgrades

We don’t talk about our infrastructure much, but Bazel is a massive part of our development methodology. It is a single system powering our build, test, and deployment infrastructure across multiple repositories and many languages. Our core repositories are – and will continue to be – built in Bazel. Its hermetic builds, global caching, and infra-as-code approach suit us perfectly.

Be aware: it has a high setup cost and is best suited for environments where Bazel is well-explored (like C or Java).** That being said, we paid those setup costs years ago and have used an untouched Bazel 6 system for the last two or three years without major issues.

The bill finally came due, and we spent a few weeks this year working through the breaking changes to upgrade from Bazel 6 to 8. That’s now done, and I expect we can leave it untouched for a few more years while reaping the benefits.

** For example, to make IDE integrations work smoothly, we ended up generating Cargo build files from our Bazel build files to support both in parallel for our Rust repositories.

TypeDB Cluster Alpha

Another area coming to fruition is the Rust rewrite of TypeDB Cluster, available in TypeDB Cloud or licensed as TypeDB Enterprise. The new version is currently available in Alpha (read: unstable, and may still have breaking changes!). While Clustering was available in TypeDB 2, we’ve made fundamentally different architectural decisions for TypeDB 3.

What this means practically: TypeDB Cluster now supports operations like cluster resizing from day one. We also have clear paths to optimize and minimize disk usage that simply didn’t exist before. On top of that, we’ve drastically improved the UX for both server administration and driver usage.

Put simply, we’re proud of the new platform we’ve built to continue enhancing our Cloud-hosted and self-hosted Enterprise offerings.

Studio Upgrades

We don’t need to go into too much detail here, as we’ve recently released a series of technical blog posts (part 1, part 2, part 3) and expositions (seeing the schema and vibe querying) about the new upgrades. It’s been amazing to see that the new agent mode and data explorer pages are already some of the most-used components in TypeDB Studio!

A huge thanks to all the community members who sent us their feedback, ideas, and bug reports.

Call me old school—I prefer light mode!

The Summer ’26 Roadmap

This year, we want to establish TypeDB as the go-to option for modeling, knowledge engineering, and AI. I strongly believe that TypeDB is the best database in the world for modeling how you think – and by extension, making it incredibly easy for LLMs to pick up and use your data.

So, what have we got cooking for the next season?

We’ve planned a series of highly requested production-grade features for TypeDB’s core, as well as some prototypes of even larger changes such as vector indexing. (Note: The code below is suggestive and may still change!)

• Batch Query: A new mode to optimize querying, and especially data loading. Expect 2-5x speedups! Even better, this feature will double as a robust way to prevent query injection attacks.

given $age: integer, $name: string; 
insert $p isa person, has age == $age, has name == $name;

• Documentation and Metadata in your Schema: We want you to think of TypeQL more like code than query. Naturally, you should be able to annotate and record extra information against it. These features will allow you to do that – and be a fantastic way to give more context to LLMs working with your data model.

define 
entity uk_company 
    @doc("A company officially registered with the UK government")
    @meta("registry", "https://www.gov.uk/government/organisations/companies-house")
    @meta("frontend_icon_url", "...");

• Structs: Create custom compound value types like coordinates or custom timestamps. This introduces an entirely new way to model data in TypeDB!

define 
struct xy: 
  x value float, 
  y value float; 

struct timestamped_xy: 
  ts value datetime, 
  xy value xy; 

attribute recorded_coordinate, value timestamped_xy;

• Serial Auto-Incrementing Attributes: automatically populate unique attributes for your data.

define 
entity employee, 
  owns employee_id @serial;

• Data Manipulation Functions, such as string slicing

match 
$p isa employee, 
  has employee_id $id; 
let $without_prefix = $id[2..];

Moving the Broader Stack Forward

At the same time, the rest of the stack also is progressing:

• TypeDB Cluster to Beta & GA: Moving toward stable, tested, and highly performant production readiness.
• Studio Enhancements: A brand new graph-based data explorer that lets you explore without writing a single query, and a built-in data loader.
• Bulk Loader CLI: A new tool to bulk load CSVs into TypeDB in a highly optimized way.

We’re also planning more AI integrations and evaluations to improve TypeQL generation and enhance TypeDB’s interoperability with the exciting work happening in the agentic world.

To further push that goal, we’ll be prototyping a vector storage and indexing component right inside TypeDB. If that’s interesting to you, please let us know on Discord or email me at joshua [at] typedb.com!

Inspired by others in our industry and some of our existing users, we’re also exploring releasing an embedded version of TypeDB. This area still requires a lot of design work, so as with vector support, please let us know what you’d like to see and how you’d use TypeDB in embedded mode.

Lastly, the infrastructure work never truly ends. We’ve got new security protocols and setting up new processes to complement our test suites. These will run continuous benchmarks and stability tests for TypeDB CE and Cluster to ensure our quality control remains high and you get the product you deserve with every release.

Let’s go!

With that, I hope you’re as excited as I am for what’s coming next. If you have any ideas, questions, or comments, our doors are wide open. Join us on Discord, or drop me an email.

— Joshua