Moorhold Data
Your self-hosted data stack, agent-ready.
Moorhold turns a stack of open-source data tools into one governed workspace for people and their AI agents. You declare each data source once and Moorhold wires it into every tool. AI agents propose changes, and a person approves them. It runs on infrastructure you own, and you can take it out again.
See how it works · Read the use case
Free for solo use. The Mac app needs Apple silicon and 16 GB of memory. There's no Windows app, and we aren't giving a date for one.
Early accessRelease 2026.11.4 is out, in early access. On a Mac, download the app and open it: no account needed. On a server, install it with an access key from your Moorhold account.
The problem
Self-hosted data tools don't run as one stack
Open-source data tools give you full control and no licence fees, but each one stands alone. Every tool has its own login and its own settings. The same database connection is typed into four places. Nothing connects a pipeline to the table it fills or the dashboard that reads it.
Joining them up by hand is real work, and it doesn't stay done. Every new source or person means another round of settings to keep in step, and nothing records who changed what.
The promise
The open-source stack, run as one
Moorhold is the layer between your people and the tools. It does not replace the query engine, the scheduler, the dashboards or the catalogue. It signs people in, wires each tool to the others through their own interfaces, keeps them in the state you declared, and records what changed and why.
AI agents are first-class users of the same workspace, with the same rules as the person they act for.
What you get
What Moorhold does
One account
People sign in through Keycloak, and each tool keeps its own permissions, with your groups projected into them.
Read more about one account
Sources declared once
Declare a data source in one place. Moorhold wires it into the query engine, the scheduler, the dashboards and the catalogue, and keeps them in step.
Read more about sources declared once
Agents propose, people approve
AI agents act for a named person, under that person's permissions. Their changes are proposals that someone else approves.
Read more about agents propose, people approve
Lineage and audit
Pipelines, tables and dashboards are joined up in one lineage graph. Every change says who asked, who approved and what it did.
Read more about lineage and audit
Removable
Export everything as plain configuration. Moorhold is designed so the tools keep working if you remove it, because it doesn't fork or wrap them.
Read more about removable
Runs on infrastructure you own
Your servers, your network, your secrets. A source's password stays in a file on your server that only root can read; Moorhold holds only the reference.
Read more about runs on infrastructure you own
How it works
Declare it, review the plan, approve it
Run the stack on your own infrastructure
Moorhold installs the open-source tools on a server you control and connects them to one sign-in.
Declare what you want
Describe a data source once: where it is, which file on your server holds its password, and which tools use it.
Review the plan
Moorhold shows exactly what it will change in each tool before it changes anything.
Approve, and Moorhold applies it
Once approved, the change is applied. If someone changes a setting by hand afterwards, Moorhold reports the difference so you can see it.
Add the source crm. The API key is read from file:///etc/moorhold/secrets/CRM_API_KEY; Moorhold keeps only the reference.
- Add: Trino schema warehouse.crm
- Add: Airflow DAG moorhold_crm__load.py
- Add: Superset datasets crm.contacts
- Add: DataHub ingestion source crm
Who it is for
People who run their own data stack
One person with a Mac
Download the app and Moorhold runs privately on your Mac, with Airflow and Superset in tabs. No server and no account needed.
Self-hosters
Teams that run their data tools on a server of their own, in an office, at home or with a cloud provider they choose. Data-residency rules are one reason to do that.
Teams moving off Microsoft Fabric
Data teams that want one workspace for loading, querying, dashboards and the catalogue, on their own server and without a Fabric capacity.
Platform and data engineers
Who run the stack and are tired of wiring the same connection into four tools by hand.
Data leads
Who approve changes and need to see who asked for what, and what it touched.
Analysts
Who want one place to start, and dashboards that keep working.
It's also a way to learn how a data pipeline fits together from start to finish, from loading a source to the dashboard that reads it.
Moorhold's desktop app is for Macs with Apple silicon, and you can download it from moorhold.dev/download. There's no Windows app, and we aren't giving a date for one. On any other computer, run Moorhold on a server and use it in your web browser.
Built on open source
The tools stay yours
Moorhold itself is proprietary software from Allotment Technology Ltd. The tools it runs are established open-source projects, used unmodified. You can upgrade them, swap them or keep them without Moorhold. Each Moorhold release pins every tool to an exact version and is signed, so a tool you upgrade yourself leaves Moorhold's release line.
| Tool | What it does | Licence |
|---|---|---|
| Keycloak | Sign-in (OpenID Connect) | Apache-2.0 (Keycloak licence) |
| Trino | Query engine | Apache-2.0 (Trino licence) |
| Apache Airflow | Scheduling and pipelines | Apache-2.0 (Apache Airflow licence) |
| Apache Superset | Dashboards and charts | Apache-2.0 (Apache Superset licence) |
| DataHub | Catalogue and lineage | Apache-2.0 (DataHub licence) |
| PostgreSQL | Database | PostgreSQL Licence (PostgreSQL licence) |
| dlt | Loading data from APIs and databases | Apache-2.0 (dlt licence) |
| dbt Core | Data models (SQL transformations) | Apache-2.0 (dbt Core licence) |
Licences checked by Allotment Technology, 28 September to 3 October 2026. Licences can change; each release is checked again before it is supported.
Talk to us
Free for solo use. If your organisation runs its own data stack, or wants to, tell us what you need.