Supabind
Try Now

See what's inside your data, then show it.

Diagnostics score every table before you trust it, Charts and the AI SQL Assistant turn a table into an answer in a couple of clicks, and the project structure keeps every table, transformation, chart, and query together so you can reopen exactly where you left off.

View pricingExplore the workspace
Demo · OverviewLive✦Orders — raw import94%RegionRevenueStatusWest$128,400East$94,120North—!South$76,900Revenue by region+18% west✦AI SQL Assistantselect region, sum(revenue)from orders group by region→ West leads at $128,400

One workspace for your data

Clean, analyze, visualize, and save - everything happens on the same live table.

Know your data's quality before you trust it.

Every table gets an overall quality score plus a breakdown across completeness, uniqueness, and consistency. Switch between Overview, Quality, Columns, and Explorer to see the same table from a different angle - a top-line score, a metric breakdown, a per-column card grid, or a single-column deep dive with its full top-values list.

In practice

Example: a 500-row, 5-column export scores 69.4% overall. The Columns tab shows Date sitting at just 30.6% uniqueness with 153 distinct values and zero nulls - Explorer then breaks that single column down further with its top recurring dates.

69%QUALITYCOLUMN METRICSCompletenessUniquenessConsistency153 distinct0 nullsExplorer open

Narrow any table down without writing a query.

Stack conditions on any column and combine them with AND/OR logic. The Data preview updates immediately as you add or adjust a filter, and the same conditions carry through to any chart or query you build from that table - Clear All resets everything in one click.

In practice

Example: filtering product_file_jun3 to Name contains "Jacket" AND PIM published equals TRUE narrows a 100-row preview down to 32 matching rows instantly.

Data FiltersClear AllANDORNameContainsJacketPIM publishedEqualsTRUEApply Filters32 of 100 rows match

Reshape columns and add computed ones, live.

Rename, retype, or reorder any column straight from the Columns panel. Add a computational column built from an existing field, a fixed value, or a formula, and see a live preview of the result before it's saved to the table.

In practice

Example: adding a computational column named "discounted_price", typed as Number and copied from the existing Amount column, previews its values immediately - no separate export required.

Columns (5)NametextQuantitytextDatedatePricenumAmountnumAdd Computational ColumnColumn namediscounted_priceTypeNumberCopy from columnAmountLive preview

Bring two tables together on a shared key.

Pick a base table, the table to merge in, and the column that joins them - the result becomes a new working table in the project, ready for its own diagnostics, charts, or AI SQL queries just like any table you uploaded directly.

In practice

Example: merging product_file_jun3 with a pricing_ref table on SKU produces a single joined table that keeps every column from both sides, one row per SKU.

product_fileSKUNameBrandpricing_refSKUPriceStockon SKUJoined table - 5 columns, one row per SKU

Ask for a chart, get one built on your live table.

Pick a table, choose X and Y fields, a chart type, and an aggregation mode - or hand the table to the AI chart assistant and let it propose the most useful view on its own. Line, area, bar, and pie all pull straight from your working data, so a chart always reflects the current step in the pipeline, not a stale export.

In practice

Example: ask the AI chart assistant to "create the most useful chart for this table" against dummy_product_data_500 - it returns an area chart plotting Quantity by Product Name, with Type, Mode, Table, X axis, and Y axis all left editable underneath.

AI CHARTType: AreaTableX axisY axis

Describe the query, get working SQL against your table.

Ask in plain language and the assistant writes the SQL for you - Create New for a fresh query, or Modify Current to adjust the one you're already looking at. It only ever runs against tables you've added to the project, shown right underneath the prompt box so you always know what it can see.

In practice

Example: with dummy_product_data_500 selected (5/5 columns, 500 rows) in Available Tables, prompting "Show me the top 10 products by sales" returns a runnable query against that table, powered by Groq's llama-3.3-70b-versatile model.

✦AI SQL AssistantGroq · llama-3.3+ Create NewModify CurrentAsk me to generate SQL queries...e.g. "Show me the top 10 productsby sales"GenerateAVAILABLE TABLESdummy_product_data_500

Every table, transformation, and export in one place.

A project keeps Source Files, Reference Files, Templates, Output, Syncs, and Workflows together in one Tables panel, with an Activity & Logs folder tracking everything that's happened in the project. Save the whole shape as a template, flip on Dev Mode to inspect the underlying query, or delete the project outright when you're done with it.

In practice

Example: reopen the "Demo" project and its seven folders are all there - Source Files through Workflows plus Activity & Logs - alongside live counts for items in project, files linked, records, and storage.

Projects›Demo0in project0files linked0records0BstorageSource Files0 itemsReference Files0 itemsTemplates0 itemsOutput0 itemsSyncs0 itemsWorkflows0 itemsActivity & Logs1 event

From raw file to finished answer.

Add a table, run diagnostics, transform or merge, build charts or ask questions, then save the project.

Bring in your source

Add a table to the project. It lands in the Tables panel with row and column counts visible immediately.

Run Diagnostics

Check the overall quality score, then drill into Columns or Explorer to see nulls, distinct values, and top values for any field that looks off.

Build a chart

Pick a table, an X and Y field, and a chart type - or ask the AI chart assistant to generate one for you - then save it to a dashboard.

Ask the AI SQL Assistant

Describe what you need in plain language and get a runnable query back, scoped to the tables already in your project.

Save the project

Keep the tables, steps, charts, and queries together. Save as a template to reuse the same structure on the next file that comes in.

FAQ

What does the Diagnostics quality score actually measure?

It blends column-level completeness, uniqueness, and consistency into one overall score, then breaks each metric out per column so you can see exactly which field is pulling the score down.

Can I chart any table in my project, or only the final output?

Any table currently in your Tables panel. Point the chart at the raw source, a mid-pipeline step, or the final output - whichever reflects what you want to show.

How does the AI chart assistant decide what to build?

It reads the table's columns and types, then proposes a chart type, X/Y fields, and aggregation. You can accept it as-is or open the same controls - type, mode, table, axes, style - to adjust it yourself.

What's the difference between Create New and Modify Current in the AI SQL Assistant?

Create New writes a fresh query from your prompt. Modify Current takes the query you're already looking at and changes it based on what you ask for, so you don't have to restate the whole thing.

Which tables can the AI SQL Assistant see?

Only the tables you've added to that project - they're listed under Available Tables right below the prompt box, so you always know the scope before you generate anything.

What's the difference between the project and a workflow?

A project holds the tables, transformation steps, diagnostics, charts, and queries for one piece of work - like the "Demo" example. A workflow is a saved sequence you can rerun on a new file without rebuilding it.

Can I save a chart to a dashboard, not just view it once?

Yes. Give the chart a title, generate or configure it, and save it - it stays on the dashboard alongside any other charts you've built for that project.