We design the tools the Qlik ecosystem is built with.

No — Q-Expert generates real, native Qlik syntax (LOAD, JOIN, CROSSTABLE, ApplyMap, and more) from the canvas you design. You can always inspect and edit the generated script; nothing is hidden behind a proprietary format.

QVD, Excel, CSV and TXT files, and sources already materialized in the Transformation layer. It deliberately doesn't connect directly to databases (SQL/ODBC) — that's a step earlier in the data architecture, handled before Q-Expert.

An interactive walkthrough that builds a real pipeline step by step — RAW, Rename, TXF, Expression Builder, Key Builder, JOIN, BUSINESS, Generate Syntax — opening the app's real windows at each step, at your own pace.

The Library panel (left, registers source files), the Canvas (center, where the model is built), the Properties Panel (right, configures the selected node), and two toolbars on top — one for project/account-level actions, one for creating nodes and controlling the canvas view.

Project-level actions: New/Open/Save/Duplicate Project, Variables, Project Settings, Export/Import as .qep, Reverse Engineering, Generate Syntax, the Routine Catalog, app Settings, and Report a bug.

Buttons to create each node type (Entities, Operations, Flow Control — each with its own color), plus Find (Ctrl+K), Duplicate, Script Split, Rename, Trace, Qualify/Unqualify, Zoom controls, Auto Layout, and a Minimap for large projects.

Go to Settings → Demos → Start Tutorial. It opens a dedicated practice project and walks you through a real pipeline — RAW, Rename, a TXF, Expression Builder, Key Builder, JOIN, BUSINESS, Generate Syntax — spotlighting exactly where to click in the real app at your own pace, with Next/Back and Pause whenever you want to explore freely.

14-day full trial, no card required: everything unlocked. After that you keep the canvas, every node type, file previews, the editors and your cloud projects for free; but the generated script (Generate Script, node preview, Deploy and Script Split) only shows the first 5 nodes, and Reverse Engineering and project export/import (.qep) are disabled.

Create your free Q-Expert account first. To upgrade to Premium, use the Upgrade plan button inside the app (About → Upgrade plan) — it's tied to your logged-in account, no email matching needed. To buy a Team pack, use the checkout link on this page with the same email as your Q-Expert account so it activates automatically.

Each teammate creates their own free Q-Expert account, then you invite them by email from inside the app (About → Team Seats). From there you can see who's using each seat and free one up to reassign it to someone else — one invoice, in your name, for the whole team.

Report a bug or an idea from the 🐛 button in the app, document it well (steps to reproduce and evidence help the most), and our team awards credits based on complexity and impact — from 10 credits (Bronze, 1 month) up to 80+ (Diamond, 2 years) of a free Q-Expert Full license; the 1-year and 2-year rewards also include a free Q-Explorator license for as long as the reward lasts. Open to any signed-in user, Free plan included, until January 31, 2027.

A Raw Script block has nothing to do with the RAW layer of the data model (RAW/TXF/BUS). It only shows up in Reverse Engineering: when it rebuilds a canvas from an existing .qvs script, anything it can't resolve with certainty is kept as a Raw Script block instead of being guessed at — the original code stays exactly as written, attached to the canvas, so nothing is lost or silently altered.

It's where the real work happens: JOIN, CROSSTABLE, ApplyMap, SK/CK keys, WHERE, GROUP BY, calculated expressions — each TXF resolves one piece of the model, in isolation. This is the layer Q-Expert is built for.

The final result, ready for the app's associative model — stored to QVD with STORE, and the only thing the BI team needs to consume.

Surrogate (SK) and composite (CK) keys. Mark a column as one and Q-Expert cascades the prefix, name resolution and QUALIFY automatically, instead of you writing that boilerplate by hand in every table.

It converts a wide table (one column per month, for example) into a tall attribute/value table — the visual equivalent of Qlik's CROSSTABLE LOAD. You set how many leading columns stay as qualifiers and name the resulting attribute and value columns.

It loads a key→value table for later use with ApplyMap() anywhere else in the canvas — the visual equivalent of Qlik's MAPPING LOAD.

KEY generates a synthetic key (SK) from one or more columns, normally built with the Key Builder to avoid Qlik's own, less readable, automatic synthetic keys when two tables share more than one column.

It repeats a whole sub-canvas once per item in a list — files in a folder, values of a field, or a numeric range — the visual equivalent of FOR EACH…NEXT or FOR i = 1 TO n. Double-click (or Enter Loop) to edit its body as its own nested canvas.

It branches the flow into different paths based on a condition's value — the visual equivalent of an IF…ELSEIF…ELSEIF…END IF with more than two outcomes, or a CASE block. Each branch can pull from a completely different source as long as they all produce the same column structure downstream.

It invokes a Qlik subroutine already defined in the Routine Catalog, passing whatever parameters that routine declares — the visual equivalent of CALL RoutineName(param1, param2). Useful for logic reused across a project, or across projects via a shared routine library.

They declare or assign a Qlik variable at that point in the flow — LET evaluates the expression before assigning it, SET assigns it as-is, unevaluated.

A data table typed directly into the canvas, with no source file — for small catalogs, mapping lists, or reference data that doesn't need its own QVD. It's also where Q-Explorator's "Inline for Q-Expert" option pastes a copied grid selection, already in the right format.

It generates a sequence of synthetic rows — typically a date or number range — with no dependency on any external file. It's the usual starting point for a Master Calendar.

It automatically generates a complete master calendar — an AUTOGENERATE date range plus the usual derived columns (Year, Month, Quarter, Week, Month name…) — without building that logic node by node.

Checkpoints are STORE / EXIT SCRIPT points with a global on/off switch, so you can enable or disable them all at once. The Script Splitter divides a long flow into several physical .qvs files, so the script stays manageable as it grows.

You switch between Dev, QA and Prod without touching the diagram — each profile keeps its own lib:// path mapping, and Deploy writes the .qvs straight to disk for whichever environment is active.

A tool for building Qlik expressions with a function search, nestable cards and a live preview, without leaving the column you're editing.

JOIN combines two tables by shared key columns (LEFT, INNER, OUTER); CONCATENATE stacks two tables with the same structure on top of each other. Q-Expert models CONCATENATE as a variant of the same JOIN node — you pick it instead of LEFT/INNER/OUTER in the Properties panel.

It converts the whole canvas — every node and connection — into the final Qlik LOAD/STORE script, ready to copy or download.

A complete project — every node, connection and setting — saves as a single .qep file. Export serializes the active project with a date stamp in the filename; Import adds a .qep to your project list, automatically renaming it if a project with the same name already exists.

Yes. Q-Expert tracks each column by a stable internal reference, independent of its visible name — so an expression, a key or a JOIN already built on that column keeps working after you rename it.

By default a JOIN/CONCATENATE's output has no identity of its own — it's absorbed directly into the downstream TXF. Check "Name the merged result" if you need that merge to be a real, named table in the generated script, for example because several nodes need to read it by RESIDENT.

Embed writes that input's LOAD literally inside the JOIN/CONCATENATE, with no separate table — only available when the input feeds nothing else. Insert filtered copy creates a real, separate TXF node (auto-embedded) for when you need a modified version of an input — a different WHERE, fewer columns — without touching the original table, which keeps working unchanged everywhere else.

Yes — "Inputs of this CONCATENATE" is the exact same mechanism as JOIN's: the same embed checkboxes, the same Insert filtered copy, Embed all eligible and Own table for all. The only real difference is CONCATENATE has no "Name the merged result" option, since it doesn't merge columns.

Yes. An Insert filtered copy is a genuine TXF node internally, so its WHERE and calculated columns support $(variable) the same as any other TXF. Inside a Loop body, the FOREACH's control variable and any LET/SET declared in that same body are available there automatically, with no extra setup.