Start free
Features

Everything you need to verify.

Learn on guided challenges. Practice on working semantic models. Verify what ran, diagnose what failed, and understand why the measure behaves that way.

Start practicing free Compare plans
01 — Editor

Write DAX like code, not like a formula bar.

Autocomplete, hover docs, syntax highlighting, and inline execution results make practice fast. Saved acceptance tests separately check the business rules you define; a query that merely runs is never described as correct.

  • Autocomplete, hover docs, formatting, and syntax highlighting
  • Code formatter and syntax highlighting
  • Execution evidence plus user-defined acceptance tests
  • Model inspector built in
  • Keyboard shortcuts for power users
Total Revenue · Retail Model
MEDIUM
1
2
3
4
-- Filtered total, context-safe
CALCULATE(
  SUMX(Sales, Sales[Qty] * Sales[Price]),
  KEEPFILTERS(Product[Category] = "Bikes")
)
● 4 of 4 test cases passing
02 — Models

Hosted practice models. Real relationship patterns.

Finance, Retail, HR, Marketing, Manufacturing, Healthcare, Education, SaaS — built with role-playing dimensions, inactive relationships, and the kind of ambiguity that separates a star-schema toy from production.

  • 2 built-in practice models, with custom uploads on paid plans
  • Role-playing dimensions and inactive relationships
  • Production-shaped ambiguity, not a toy schema
  • Browse by topic, difficulty, or model
01
Finance
02
Retail
03
HR
04
Marketing
05
Manufacturing
06
Healthcare
07
Education
08
SaaS
03 — Daxie

An AI tutor that teaches the eye, not just the answer.

Daxie can use your current code, model metadata, last error, and a bounded result preview to explain what happened. Result rows are excluded by default and require explicit confirmation before they are attached to a message.

  • Model-aware: reads bounded metadata, current code, and last error
  • Result-aware only after preview and explicit consent
  • Attachment limits and redaction are shown before sending
  • Teaches concepts, doesn’t hand over solutions
  • On-demand hints per challenge
  • Available on every plan, including free
🦊
Daxie
Your DAX tutor
ONLINE
Why does my CALCULATE return blank?
Good question. Look at what your filter argument is doing to the row you’re measuring. Try wrapping it in VALUES('Date'[Year]) — what changes?
04 — Courses & Progress

Curated paths. Personal telemetry.

Six free courses, fundamentals through advanced, all against real AdventureWorks data. Plus a dashboard that tracks what you’ve mastered and where you’re weakest — so you know where the silent failures will catch you.

  • Six free guided courses, 300+ problems total
  • Fundamentals → Time Intelligence → Filter Context → Advanced
  • Per-topic strengths and gaps
  • Streaks, personal bests, community percentile
SOLVED
147
of 300+ problems
STREAK
12 days
personal best
MASTERY BY TOPIC
Filter context82%
Time intelligence64%
Text functions38%
05 — Verify & Acceptance Tests

First prove it ran. Then prove it met your rule.

Verify creates an honest execution receipt with the target, context, result, and limitations. Acceptance tests make the separate business claim explicit by checking expectations you define.

  • Receipts preserve source, model, filters, outcome, and limitations
  • Check expected scalars, ranges, blanks, and direction of change
  • Save reusable suites and retain verification history
  • A successful query is never mislabeled as business-correct
See how Verify works
VERIFICATION RECEIPT
Execution proved
Expected total rangePASS
Previous period directionPASS
Blank handlingPASS
06 — Agentic Debug

Turn a failing measure into an evidence-backed diagnosis.

Debug inspects the code, error, model metadata, and bounded probes in a visible sequence. It stops within a defined budget and shows the evidence behind its diagnosis.

  • Syntax, relationship, filter-context, and result probes
  • Visible run steps, tool budgets, and stopping conditions
  • Suggested changes remain advisory until you run them
  • Share a safe diagnostic summary without exposing private data
Explore Agentic Debug
  1. 01   Parse the failure
  2. 02   Inspect model context
  3. 03   Run bounded probes
  4. 04   Explain the evidence
07 — Visual Explain

See filter context move through a measure.

Explain turns supported DAX into a step-by-step visual walkthrough of row context, filter context, relationships, and evaluation. Partial support and assumptions stay visible.

  • Trace CALCULATE and iterator context transitions
  • Connect each step to the relevant table relationships
  • Keep unsupported constructs and assumptions explicit
  • Use public demos free, then explain measures on uploaded models
Watch an Explain walkthrough
ROW CONTEXT
Sales iterates one row at a time
↓ CALCULATE ↓
FILTER CONTEXT
The current row becomes active filters
08 — Power BI · Read-only

Run the proof loop against an existing semantic model.

Professional users can connect with Microsoft, choose an eligible workspace and dataset, and execute read-only queries. DAX Solver does not publish model changes back to Power BI.

  • Microsoft identity and explicit dataset selection
  • Encrypted server-side credentials and revocable access
  • Verify, Explain, Debug, and acceptance-test connected measures
  • Structural fixes stay advisory; the source model is unchanged
Compare Professional
CONNECTED SOURCE
Power BI semantic model
READQUERYVERIFYNO PUBLISH
Individual

Move from examples to your own model.

Upload a model you are authorized to use, ask Daxie about bounded results with explicit consent, and keep verification history and acceptance tests tied to the model.

  • Five uploads, up to 5 MB each
  • Result-aware guidance with preview and consent
  • Saved verification receipts and acceptance tests
  • Higher Daxie and Debug allowances
Professional

Repeat the diagnose-and-prove loop on work models.

Professional raises model, history, and agent allowances and connects to eligible Power BI datasets read-only.

  • Twelve uploads, up to 16 MB each
  • Twenty bounded agentic Debug runs per day
  • Higher acceptance-test suite and history limits
  • Read-only Power BI execution; structural rewrites stay advisory
Evidence labels

Every result names its source of truth.

A receipt is useful only when you know what data produced it. DAX Solver keeps that boundary visible, from safe practice rows to a connected Power BI dataset.

See how verification works
01
PBIT · synthetic

Practice the behavior safely.

Generated rows shaped by the template schema let you test how a measure behaves without using production data.

Receipt proves Behavior on generated rows
Not production values
02
PBIX · hosted

Test the file you brought.

Run against imported model data in a private, account-owned workflow when you have permission to process it.

Receipt proves Result on the imported file
Private imported data
03
Power BI · read-only

Check the live dataset.

Execute against an eligible connected dataset while DAX Solver keeps structural changes advisory.

Receipt proves Result on the connected dataset
Read-only · no publishing

Start free.
Scale when you need to.

Build the judgment for free. Upgrade when your own models and repeated evidence become the work.

Start practicing free Compare plans