Dashboards · Build

Say what you want to know. Get a finished board.

Start from one sentence, a template, an example you like or a report you already have. Then change anything by asking or by clicking, and describe any chart the library doesn't have.

Try free Try the live example

Live example

Including a chart no library has

The rings on this board are a custom chart. It was asked for in words: “a radial bar chart of revenue by store, one ring per store”. No standard chart draws rings, so Mammoth drew one, and every value on it still comes from a query on the data.

Sample data for a fictional coffee chain. Also on this board: an icon stat and a pictograph.

Open the full board

Start

Start from a sentence, a template or your own file

Make a board from one sentence

Pick a dataset, say what you want to know, and Mammoth builds one finished board, showing each stage as it works.

Suggestions from your own data

If you don't know what to ask, the create screen offers starting points based on the shape of your data, each naming real facts from it.

Start from a template

Browse ready-made boards, decks and documents that open on sample data, preview them full size, and adjust the look before taking one.

Use my data on a template

Upload your file to a template; Mammoth fits the template to your columns in the background, suggests your version, and shows what changed and why before applying it.

Honest template fit

Before you apply a template you see how well it fits your data, with the reasons whenever it is less than a great fit.

Public template gallery

Anyone can browse the templates at mammoth.io/templates without signing in, and the chosen template carries through sign-up.

Make it yours

Change anything, by asking or by clicking

Edit by chat or by controls

Change any chart, page or style by typing in chat or by using the controls beside the board. Chat edits stay on the item you selected.

Describe any chart

Describe a chart in words. If the chart library has it you get that chart; if not, Mammoth draws a custom one from a query over your data.

Save a custom chart for your team

Save a custom chart as a chart type for the workspace so anyone who can edit a board can add it.

Add a chart yourself

Add charts by hand, starting from a list of every column with sample values, and see live previews before you place one.

A broad chart library

Bars, lines, maps, scatters, sankeys, funnels, waterfalls, big numbers and more, each drawn the same in the browser and in exports.

Icon stat and pictograph charts

Show a headline number with an icon, or a count drawn as rows of repeated icons.

Attach a picture or brief in chat

Paste a screenshot to make a board look like it, or a brief to steer it; nothing changes until you press Apply.

Tidy up a page

One click lines up the tiles on a page you arranged by hand.

Bring what you have

Start from your existing reports and know-how

Bring a report over

Drop in a Power BI or Tableau file, a PDF or a screenshot, and Mammoth builds its own board answering the same questions.

Build from an example

Upload an example of the output you want, even a screenshot or a scan, and edit the extracted layout before Mammoth builds on your data.

Teach Mammoth your business

Save reusable background, goals, definitions, emphasis and rules that guide what the AI writes, including viewer Q&A answers; it can fill them in from a document.

Review a context change first

When a context is added, changed or removed on a board, you see what would change before it applies, and a banner flags boards whose context has moved on.

Save your board as a template

Turn your own board into a template that can be applied to other data.

Add a chart from the data editor

Send a chart from the data editor's Explore cards to a board's draft.

Limits

Limits worth knowing

Better to know before you choose than after.

AI generations are allowed per plan: One request makes one board. Each plan includes a monthly allowance of AI generations; see pricing.
Custom charts are drawings, not code: The AI writes a drawing recipe, never code or values. Projections are refused. A custom chart may not get every filter, export or style feature that library charts get.
A poor template fit gets options, not a half-empty board: One AI pass per upload. A poor fit offers options (reshape, other templates, build in this design, keep what fits) rather than a half-empty board. Only four kinds of calculated field.
Imported reports start on sample rows: Power BI and Tableau imports are built on a fitted sample, not your numbers, until you point the board at your data. The original is not reproduced pixel for pixel. Example files up to 20 MB; very large workbooks may not import.
Examples say what they couldn't match: It lists what it could not match and asks how to calculate any rate the example never defines.
Context files are read as text: With no context attached, every context in the workspace applies. Raw data tables and finished reports are refused. PDFs and Word files are read as text, and scans are not read.
Some boards can't become templates: Refused if the board can't re-fit its own data, or if it reads more than one table.

Checked against the product on 2026-10-23. Every dashboard capability, with its plans and limits, is also published as machine-readable JSON.

Questions people ask first

Do I need to know which chart to use?

No. Say what you want to know and Mammoth picks charts your data can support. If you have a chart in mind that the library doesn't have, describe it and Mammoth draws a custom one.

Can I bring a Power BI or Tableau report over?

Yes. Drop in a .pbix, .twb or .twbx file, a PDF or a screenshot, and Mammoth builds its own board answering the same questions. A Power BI or Tableau import is built on a fitted sample until you point it at your own data.

Can I edit by hand?

Yes. Every chart, page and style can be changed by typing in the chat or with the controls beside the board, and you can add charts yourself from a list of every column.

Does the AI write code for custom charts?

No. It writes a drawing recipe over an ordinary query: which shape stands for each row and which column sets its size. The recipe has no place for a value or for code, so every number still comes from your data.