For developers and ops teams running on Postgres
Skip the dashboard builder. Get an admin panel in 30 seconds.
Point OnlyDash at your database. We read the schema, sample your data, and hand you back a working back-office — tables, forms, filters, edits — for every table you have. No drag-and-drop. No templates. No setup wizard.
Postgres and Google Sheets today · MongoDB and MySQL in beta
The League table from a real 25-table betting analytics database, generated without one line of UI code.
Paste a connection string. See it for yourself.
Read-only until you say otherwise. Disconnect any time.
Try it now
Paste a database connection string. We'll spin up an admin panel for it in seconds — read-only until you say otherwise.
Other “no-code” tools still want you to build the UI
Every internal-tool platform on the market hands you a drag-and-drop canvas and a panel of widgets. You still wire up each table, bind every form field, copy SQL into action nodes, and end up with a week of work before a non-technical teammate can edit a row. OnlyDash is the inverse: the database is the design.
Three steps. No widgets.
What a fresh datasource looks like, from paste to live admin panel.
Paste a connection string
postgresql://… is all we need. We test it in front of you before saving, and the password is masked in the UI from the moment the datasource is created.
We read your schema
Tables, columns, primary keys, foreign keys, indexes. We sample up to 1000 rows per table with TABLESAMPLE SYSTEM to infer subtypes — URL, UUID, JSON, dates, enums — and you get to watch it happen.
Use the UI
Every table becomes a page with a data grid, an Edit form, a Create form, filters, sort, pagination, and one-click related-row expansion. The forms know your column types and pick the right input for each one.
What you get on day one
Not a deck of features. Just the things any back-office actually needs, generated automatically for every table in your schema.
Schema introspection
Base tables, columns, types, nullability, primary keys, and foreign keys read live from information_schema. Views and materialised views are intentionally skipped.
Type inference from real data
Each column is sampled and classified — UUID, URL, email, phone, hex colour, date, JSON, enum — so the right input shows up in forms and the right formatter in tables.
CRUD without writing endpoints
Create, edit, and delete rows on any table from the UI. We use your database's own keys; nothing is mirrored or duplicated.
Filters, sort, pagination
Every column is filterable and sortable. Filter groups combine with AND / OR. Cursor pagination keeps things fast on tables with millions of rows.
Related-row expansion
Foreign keys are detected automatically and rendered as expandable links — open a row to see the related records in a child table without leaving the page.
Fast on big schemas
We use TABLESAMPLE SYSTEM rather than offset scans, planner row estimates instead of COUNT(*), and skip subtype regex on large text blobs. Analyse runs in seconds on multi-million-row tables.
Filter chips combine with AND / OR. Every column is searchable without a single SQL query from your side.
How OnlyDash compares
| Builder tools | OnlyDash | |
|---|---|---|
| Time to a working admin panel | Days to weeks | Seconds |
| What you do per table | Add widgets, bind fields, write SQL actions | Nothing |
| When the schema changes | Edit every dashboard by hand | Re-run analyse, UI updates |
| What non-technical teammates can do | Use what you built for them | Same — and edit data straight away |
Frequently asked questions
Which databases do you support today?+
PostgreSQL and Google Sheets are fully supported. MongoDB and MySQL connectors are in beta — we run schema discovery against them but the data-table polish is not quite there yet. The connector layer is small; new sources are typically a one-week project.
How safe is my data?+
Your database stays where it is. OnlyDash holds the connection string (encrypted) and metadata about your schema — table names, column names, inferred types, sampled value statistics — in our own Postgres. We do not copy row data into our own storage; CRUD operations are made against your database in real time.
Will it work on my 50-table, multi-million-row database?+
Yes. Schema introspection reads from information_schema (cheap), and sampling uses TABLESAMPLE SYSTEM which is O(sample size), not O(table size). End-to-end analyse on a 25-table production schema with a 22M-row fact table takes about 20 seconds.
Can non-technical teammates use it?+
Yes — the generated UI is what they use. They land on the same data grid you do, click into a row, edit it, hit save. No SQL, no widget palette, no learning curve. Roles and per-row permissions are on the roadmap.
How is this different from Retool / Appsmith / Budibase?+
Those are excellent UI builders. OnlyDash is not a builder. We do not give you a canvas with widgets to drag. We give you the finished admin panel for every table in your database, generated from the schema. If you need bespoke flows and custom widgets, a builder is the right tool. If you need the standard CRUD experience without spending a sprint on it, OnlyDash is the right tool.
What happens when my schema changes?+
Re-run analyse. New columns appear in forms automatically, new tables become new pages, and existing UI updates without manual rebinding. The metadata DB keeps every analyse attempt so you can compare what changed.
What about charts and BI?+
Not yet. We deliberately did not ship half-broken charts. Automated chart suggestions and real-time-updating tiles are on the roadmap, after we are happy that the admin-panel core is rock solid.
What does it cost?+
Free while we are in early access. Once we open pricing it will be a flat fee per datasource per month, with a generous free tier for small databases. No per-seat pricing; non-technical teammates do not cost extra.
On the roadmap, not in the box yet
Automated charts, role-based access control, real-time updates, and shareable read-only links are next. If you need them today, email us and we'll tell you where in the queue they sit.