Asset Configuration
The Assets settings page lets you control how merchi.ai handles documents found during scraping. You can choose which document types to store, how large a file we will accept, and which stored types are passed to the AI model to enrich generated product content.
PDF Documents
When merchi.ai scrapes a product page it will look for PDF documents — data sheets, installation manuals, certificates, and similar files. You decide which types are stored and whether any of those stored files are used to help generate richer product content.
Document types
The table lists all document types that the scraper knows about. For each type you have two switches:
| Column | What it does |
|---|---|
| Store | When on, any PDF of this type found during a scrape is downloaded and stored in your account. Storage is billed monthly at 1 credit per GB (rounded up to the nearest GB). |
| Use in AI | When on, a stored PDF of this type is passed directly to the AI model alongside the product images. This lets the AI read the full document content — useful for data sheets, certificates, or manuals where the text carries important product detail. |
Use in AI requires Store to be on. The toggle is disabled until you enable storage for that type.
Model compatibility: PDF content is only passed to the AI when your account is using Google Gemini or Anthropic Claude. OpenAI models do not support native PDF input, so PDFs will be ignored during content generation if an OpenAI model is selected. To use PDFs in AI generation, ensure your model is set to a Gemini or Claude variant in your account settings.
Enabling large document types (such as brochures) for AI use will increase the token cost per product run.
Adding a custom document type
If your suppliers use a document type not in the default list, you can add it:
- In the Add custom document type section, enter a descriptive name (e.g. “Test Report”)
- A key is generated automatically from the name
- Once added, the type appears in the table above — enable Store and/or Use in AI as needed
The custom type is immediately recognised by the scraper on the next run. You do not need to make any manual changes to the scrape prompt.
Maximum file size
Files larger than the configured limit are skipped and not stored. The default is 50 MB. You can set any value between 1 MB and 500 MB.
Large files increase storage costs and, if enabled for AI use, increase the token cost per run. For most product documentation 50 MB is sufficient.
Storage billing
PDF storage is an enterprise feature. Usage is calculated at the end of each billing cycle:
- The total volume of stored PDFs is measured in GB
- Any partial GB is rounded up to the nearest whole GB
- 1 credit is deducted per GB charged
- If your credit balance goes negative it will be carried forward and reconciled with your account
Storage costs accumulate over time as more PDFs are scraped. If you want to reduce storage, contact your account manager and we can arrange deletion of stored files.
Viewing stored PDFs
PDFs stored for a product are visible in the product detail modal, below the images and generated content. The table shows the file name, document type, and a link to open the file. Click the PDF icon to open the document in a new tab.
Tips
- Start with Data Sheet and Certificate enabled for both Store and Use in AI — these carry the most structured product information.
- Enable Brochure for AI use only if your products have sparse data otherwise, as brochures can be large and expensive to process.
- Leave User Guide and Installation Manual as Store-only unless the AI is missing technical detail that only the manual contains.
- Review usage on the billing page to understand how PDF storage is contributing to your monthly credit consumption.