Build visual pipelines that turn dirty spreadsheets and documents into clean, structured output, and trace exactly which step did what.
Four stages, one pipeline. Jump straight to any of them.
Pull data straight from where it lives. Push results to where your team already works.
Design a flow in the visual builder, publish it, and anyone can run it as a one-click converter. No account needed.
Real data is unpredictable. The test adapter captures the breakage, and the AI assistant rewrites the pattern until the test passes again.
Capital of France?;ParisWhat is DNA?;;Deoxyribonucleic acidLargest planet?;JupiterSpeed of light?|299792458 m/s/^(.+?);(.+)$/gmWatch a folder or run on a schedule. Every new file is transformed and delivered straight to your destination.
Every feature exists to make your text transformations traceable, testable, and production-ready
Trace exactly which node in your chain transforms A into B. Every step shows its input, output, and logic, making continuous improvement intuitive, not guesswork.
Apply your flow to an entire folder of text files at once. All processed results are neatly organized into an output folder. No more one-file-at-a-time copy-paste.
Switch between three views of the same flow: a visual graph builder, a linear step builder, and a purpose-built text DSL. The DSL is faster and AI-friendly, while every mode keeps full analyzability. Highlight any section to see its process-level data autopsy.
Give AI an origin and destination sample and it builds the flow, then verifies it with TDD, testing that the output actually matches your expected result. Or prompt it for a single tricky pattern and keep the logic yours. Reliable transformations without trusting AI to do the work directly.
Start free. Upgrade when your workflows outgrow it.
Build your own flows and run shared ones.
Unlimited flows, automation, and destinations.
Shared workspace for departments.
Everything you need to know before you start building.
A visual flow builder that turns messy, semi-structured content - spreadsheets, exported docs, pasted text - into clean, import-ready formats like Moodle XML, Aiken, and GIFT. You build a transformation pipeline by dragging and connecting nodes.
No. You build flows by dragging nodes onto a canvas, and shared flows run with a single click. Regex is there when you want fine-grained control, but it is never required to use a flow someone else built.
Yes. Flows execute entirely in your browser - your content is not uploaded to a server to be processed. That privacy model is the same whether you build a flow or run a shared one.
Virtually any text-based format. Because flows are regex-based, you can shape output to almost any structure you need - there is no fixed list. Quiz and question formats for learning platforms (Moodle XML, Aiken, GIFT, QTI) and general formats like CSV, JSON, or custom markup are just common examples, not limits.
Yes. Image handling is built in, so questions and content keep their figures instead of being stripped - something most conversion tools cannot do. Image support is still an experimental feature, so expect rough edges as we refine it. Image handling is available on every plan.
Yes. Publish a flow to a link and others can run it one-click on their own data, with no account setup or knowledge of the internals. Power users can fork a shared flow and tweak it for their own format variations.
Stop guessing why your output looks wrong. Start tracing, improving, and scaling your text processing with full visibility.