Test With Your Own Receipts

The best way to evaluate Taggun is on your own documents — the receipts and invoices your users will actually submit. Test single receipts in your browser with the API playground — no code required.

Test in the API Playground (No Code)

Open Upload File (verbose) in the API Reference.
Paste your API key into the apikey header field. (No key yet? See the Quick Start.)
Attach the receipt or invoice file.
Click Try It, inspect the returned JSON response, and save the result with the original file.

👍
Use the verbose endpoint when evaluating
The verbose endpoint uses the broader extraction response schema and can return detected fields and enabled feature envelopes when available. Optional properties may be omitted; fields explicitly documented as nullable may be null, and arrays or objects may be empty where their schema permits. The simple endpoint is for production flows that only need core fields.

What to Put in Your Test Set

Start with a small smoke set to check authentication, request shape, and obvious field coverage. For an accuracy decision, use a labelled set that represents your production countries, merchants, document types, image quality, and edge cases. Size the set for the decision and variation you need to measure; a fixed document count is not a guarantee of statistical adequacy. Contact us for a structured pilot at larger volume.

Real receipts and invoices — not stock photos or perfect scans.
Your countries and languages — see Supported Countries & Languages.
Your document types — receipts, tax invoices, and digital receipts representative of your production workflow. See Document Types & File Requirements.
Typical capture conditions — include the devices, lighting, image quality, and document condition your users actually produce.

Evaluate the Results

Field coverage — check the fields you need against Data Fields That Taggun Extracts.
Understand the response — see Understanding Your Results.
Report problems — contact support with the email attached to your account, the affected trackingId when returned, the sample document when safe to share, and a clear description of the issue. See Contacting Support.

When to Talk to Us

Testing at a volume the playground can't handle? We run structured pilots — batch runs over your historical data with results reviewed together — for volume and enterprise use cases. And if accuracy on your document types isn't where you need it, that's a conversation, not a dead end: performance tuning for specific document types is available on Advanced plans and above. See Plans & Feature Access.

Need something not covered here? Contact [email protected] — see Contacting Support.