MindsHub
Star — Log in Get started

About the rebrand

MindsHub MindsDB

How the product evolved from in-database ML to LLMs to open-source AI agents — and why MindsDB rebranded to MindsHub. Same company, same mission, clearer name.

  1. 2018 → 2022

    Bring AI to the data.

    MindsDB started with one clear pivot: instead of forcing teams to move their data to wherever the model lived, bring the model to wherever the data already lived. Inside the database. With the same SQL ergonomics the team already used. No exports, no separate ML stack, no specialist hire required to ship a first prediction.

    The product fit was natural enough that the name made the architecture obvious: the work happened next to the database, so the suffix was DB. At the time, most useful ML data was sitting in databases — Postgres, MySQL, MongoDB, Snowflake — so the project lived there too.

  2. 2023 → 2024

    LLMs arrive; we get broader.

    When large language models broke open the kind of work AI could do, the constraint stopped being "fit a model to tabular columns" and became "ground a model in the entire heterogeneous mess of an organization's information." Structured rows in a warehouse, unstructured PDFs and emails and Slack messages, semi-structured JSON, real-time event streams — all of it became fair game for AI to read.

    MindsDB broadened to match. The same in-the-data principle now had to handle text, documents, vectors, and the kinds of long contexts LLMs need. Knowledge bases, semantic search, and a federated query layer let agents read from many sources as if they were one. The DB part of the name stayed; the surface area was no longer just ‘the database.’

  3. 2025

    Open-source agents become first-class.

    By 2025 the most interesting AI work wasn't 'add a model to a query' — it was agents that plan, call tools, run code, and ship finished work. The frontier labs shipped capable proprietary agents quickly. The open-source community built its own. What was missing was a place to run open agents in production with the same operational guarantees the closed offerings had.

    We built MindsHub for that. A platform where open agents — OpenClaw, OpenClaw + GBrain, NanoClaw, Anton, Hermes, and more — get model routing and a hosted runtime today, with a unified credentials vault, cross-agent memory, and audit-able scratchpads coming soon. First-class. The same operational backbone the proprietary offerings keep behind their own walls.

  4. 2026 — the rebrand

    MindsDB → MindsHub.

    The product surface had moved well past 'in-database ML.' We were a platform for running open AI agents that happen to be very good at reaching the data and tools those agents need. New users hit the name first and kept asking the wrong question: is this a database?

    It wasn't. It hadn't been for a while. So we renamed the platform to match what it actually does — MindsHub: a hub for open agents — and kept the parent identity as MindsDB. The mission, the team, the company, the commitment to open source: all the same. Same minds. New shape.

The name, in one line

Same Minds. New shape.

2018 MindsDB Bring AI to the data.
2026 MindsHub A hub for open AI agents.

MindsHub is by MindsDB. The company didn’t pivot; the product surface did. Read the full company story on /about.

FAQ

Common questions about the rebrand.

Did the company change hands or pivot?
No. Same company, same team, same investors. MindsHub is the new product surface; MindsDB is the parent identity. The mission to democratize access to AI is unchanged.
Is the MindsDB query engine going away?
No. The query engine and the in-data integrations continue as part of the MindsHub platform — they're part of how agents reach the systems where data lives.
Do old mindsdb.com URLs still work?
Yes. Every path on mindsdb.com 301-redirects to the matching path on mindshub.ai. The redirect is in place indefinitely.
What does this change for existing MindsDB users?
Nothing about your existing workloads needs to move. The open-source MindsDB project continues. If you want to try open agents on the same platform — OpenClaw, OpenClaw + GBrain, NanoClaw, Anton, Hermes — they're available under the new MindsHub surface.
Why was the old name confusing?
The 'DB' suffix made sense when the work happened inside databases. Once the product handled documents, vectors, agent runtimes, and tool-using workflows, the name stopped describing what the platform actually does — and new users kept asking whether MindsDB was itself a database. The rebrand resolves the ambiguity.