NewWarehouse sources and scheduled reports are live

Your spreadsheets already have the answers. Dashful builds the dashboard.

Upload the exports you already live in. An AI agent profiles them, proposes a dataset, and waits for your approval — then turns it into a dashboard you can publish, schedule and trust. No SQL, no BI consultant, no six-week project.

app.dashful.com

Q3-receipts.xlsx

1.2M rows · 14 columns

Store receipts

3 sources linked

Approved

Net revenue

$4.82M

+12.4%

Shrink rate

1.8%

−0.4pt

Avg. basket

$38.10

+2.1%

Revenue by region

Last 6 months · live data

Published
NorthSouthEastWestCentralOnline

How it works

Four steps from a file on your desktop to a report your team opens every Monday

The whole path is deliberate. You see what the agent found, you approve what it proposes, and nothing reaches anyone else until you publish it.

  1. 01

    Drop in the files

    Excel or CSV, up to a gigabyte, straight from your desktop — or connect the shared drive, SharePoint folder or warehouse they already land in. Fifteen files at once is a normal Tuesday.

  2. 02

    Say what you want to know

    "Show me shrink by store against last quarter." The agent profiles every column, works out how the files relate, and comes back with a dataset — its grain, its measures, and every assumption it made.

  3. 03

    Approve what it found

    Nothing is loaded and no dashboard is built until you say yes. Approve one proposal, or approve fifteen at once and get a single dashboard across all of them.

  4. 04

    Publish it

    A published report is a frozen layout over live data at its own address. Grant it to a team or to one region, put it on a weekly schedule, and drop next month's file on top when it arrives.

What you get

Everything between a raw export and a number somebody acts on

Dashful is not a chat window bolted to a chart library. It is the whole path — loading, cleaning, relating, checking, building, publishing — with a person in the loop at the point that matters.

ApproveChange it

Datasets you approved, not datasets you hope are right

The agent does the tedious part: profiling a million rows, spotting that store_code and store_number are the same key, noticing the merged title row above your headings. Then it stops and shows you what it plans to build.

Approve it and it becomes a real table with named measures and quality checks that re-run every time the data changes. Send it back and say what it got wrong.

Reads the file, not the file name

Merged cells, title rows, a CSV saved as .xlsx, a pre-2007 workbook — all handled before you ever see a column list.

Relationships you declare

Facts at different grains stay separate. You say which columns are the same key, and a filter on one reaches the others.

Checks that keep running

Quality checks travel with the dataset and re-run on every refresh, so a bad file is caught before a meeting is.

Dashboards that give the same answer twice

Every widget is a declarative spec executed by a deterministic engine. The model never writes the SQL behind your dashboard, which is why the same question asked in March and in June returns the same number.

Click a bar to cross-filter everything linked to it. Ask the built-in analyst why a figure moved. Export any widget — or the whole dataset — as the data behind it, not a screenshot.

Cross-filtering that follows the links

Click once and every widget reacts, even the ones reading a different dataset that names the column something else.

An analyst on every dashboard

Ask why a number moved and get driver analysis over the data as currently filtered — read-only, scoped to that dashboard.

The numbers can leave

A widget, a dashboard or a whole dataset downloads as CSV or Excel, filtered exactly as the page is.

Published

Publish once, refresh forever

A draft is yours. A published report is a frozen layout over live data at its own address, branded as yours, visible to exactly the people you granted it to.

Put it on a daily, weekly or monthly schedule and what lands in an inbox is the report itself — the headline figures and the first rows, not a notification telling someone to go and look.

Frozen layout, live data

Later edits sit as unpublished changes until you publish again, so nobody's Monday report reshapes itself overnight.

One report, not all of them

Access is per report as well as per feature, so a regional manager can be granted their region and nothing more.

Scheduled with its own filters

Two schedules on one report can carry two regions' numbers, each executed at send time.

Sources

Wherever the data already lives

A source is just a place files arrive from. Connect one and new exports are picked up on their own — a new month's file lands in the dataset it continues, and everything built on it rebuilds.

Files

Excel (.xlsx, .xls) · CSV · PDF and Word as reference material

Drives and folders

OneDrive · SharePoint · Google Drive · Box · Dropbox · a watched network folder

Databases and warehouses

PostgreSQL · MySQL · BigQuery · Snowflake · GA4

Each connection uses your own OAuth app and a read-only role where one exists. Credentials go to a vault, never into a prompt, and never back to the browser.

Why teams switch

The reporting work that never made it onto a roadmap

Most teams do not need another BI platform. They need the four days a month somebody spends rebuilding the same workbook back.

No SQL

The person who knows the business builds the report

The people who understand shrink, margin and fill rate are rarely the people who write SQL. Dashful puts the build in their hands and keeps the engineering team out of the reporting queue.

Every month

Next month's file does not mean next month's rebuild

A dataset continues. Drop the new export on top and the rows append, the checks re-run, and every report refreshes — instead of somebody re-pasting into a workbook and re-pointing the charts.

One number

Everyone is looking at the same figure

Measures are named and defined once, on the dataset. A published report is a single address, not a workbook forwarded four times and edited twice.

Pricing

Pricing

Start free on your own files. Move up when the whole team is opening your reports.

Free

Load a few real files and build your first dashboard, with no card and no deadline.

$0
  • Excel and CSV uploads
  • Build dashboards with AI
  • One seat

Starter

For the person who owns the monthly reporting and wants it to stop being manual.

$49 / month
  • 3 seats
  • 2M AI tokens a month
  • 5M rows loaded a month
  • 10 datasets
  • Reports: 10
  • Publish and schedule reports
Most popular

Team

For a team running reporting across departments, straight from the systems the data lives in.

$249 / month
  • 15 seats
  • 20M AI tokens a month
  • 100M rows loaded a month
  • 100 datasets
  • Reports: 100
  • Drive, database and warehouse sources
  • Publish and schedule reports

Free

Load a few real files and build your first dashboard, with no card and no deadline.

$0
  • 1 seats
  • 100K AI tokens a month
  • 100K rows loaded a month
  • 2 datasets
  • Reports: 1
  • Publish and schedule reports

Enterprise

For running Dashful inside your own network, on your own identity provider and terms.

Custom
  • Self-hosted or private deployment
  • Your OIDC provider and directory groups
  • Volume limits set with you
  • Dedicated support and onboarding

Questions

Frequently asked questions

The things people ask before they hand over their first spreadsheet.

Get started

Bring your worst spreadsheet

Upload the file you dread opening and see what Dashful makes of it. Free to start, no card, and your first dashboard in an afternoon.

Keep an eye on what we build

Occasional notes on reporting, messy data, and what we shipped. No more than once a month.