Product demo · 109 seconds
From supplier PDF to your PIM, automatically
A sample supplier datasheet, extracted and audited into structured JSON, then pushed into a live PIM as typed, unit-aware attributes by a short script over the PIM’s REST API. Recorded end to end, silent, with captions.
What you are actually looking at
- Sample documents, real pipeline
- The five datasheets are samples we built to read like the ones suppliers send, including a German-language transmitter TDS, with fictional manufacturers so they can be published. Each ran through the same pipeline any upload goes through: extraction, audit, and structuring. On the battery datasheet in this recording, 24 of its 25 extracted fields landed as typed PIM attributes.
- Recorded as a script; now built into the product
- In this recording the push is a short script calling the PIM’s standard REST API, against a self-hosted Akeneo Community Edition instance. Since then, Akeneo sync has become a built-in SpecMake feature: preview the field mapping against your own attribute set, approve it, and push from the product page. SpecMake is not trying to replace your PIM; it delivers the structured, audited, source-anchored data, through the app, the built-in sync, or our REST API, so it lands wherever your catalog lives.
- Units and provenance survive the trip
- Measurements arrive as typed attributes with the unit in the attribute rather than buried in a string, and every record keeps the source file, revision, detected domain, and language it came from. That is what makes the data auditable after it lands. More on how the pipeline gets there in how SpecMake works and the extraction feature page.
What happens, scene by scene
The complete caption text from the recording, in order.
- 1
Title card
From supplier PDF to your PIM, automatically. A real workflow, recorded end to end on a live Akeneo CE instance.
- 2
The source document
The input: a sample supplier datasheet, a 48 V battery module, as a PDF.
- 3
Structured output
SpecMake extracts and audits it into structured JSON: typed values with units, confidence scores and source quotes.
- 4
The push
A short script pushes that JSON into the PIM through its standard REST API.
- 5
The catalog
Five datasheets became four products and one product model, families assigned, 100% complete.
- 6
Typed attributes
Values arrive typed: 51.20 Volt, 100 Ampere hour. Units live in the attribute, not in a string.
- 7
One document, three sizes
One datasheet described three valve sizes side by side.
- 8
Product model and variants
So it became a product model plus three variants: shared values live once on the parent, each variant carries only what differs.
- 9
Provenance
Every record keeps its provenance: source file, revision, detected domain, and language.
- 10
Closing card
The data preparation layer for your PIM. Extract, audit, structure, then push to the platform you already use.
See it on one of your own datasheets
Run one of your own spec sheets: the fields come back structured, with the compliance gaps per EU regulation and a DPP readiness score when the product category is in scope. The file is not stored, and no account is required.
The Charter Customer Program is a small cohort of manufacturers putting the pipeline to work on real product lines, at 25% off for six months.