Files

Upload

Click Upload, then drop your files into the dialog or click Browse.... You can upload several files at once. File names must be unique within your account.

File limits

• Maximum file size: 1000 MB

• Maximum text length per record or TXT file: 2000000 characters

File Preparation Guidelines

To ensure successful upload and processing, please follow these requirements when preparing your files:

1. Remove all non-table content

• Delete any titles, notes, comments, customer info, footers and empty rows.

• The file must start with a single header row.

• Each column must have only one header cell (no merged header fields).

• Do not include multiple tables in one file.

2. Ensure a clean table structure

• One row = one item/record.

• No merged cells anywhere.

• No separate tables side by side or below each other.

• All rows must follow the same structure (same number of columns).

3. Use plain text only

• Remove all images, icons, emojis, bullets, shapes, hyperlinks or embedded objects.

• If an image was originally present, delete it or replace it with text (e.g., link or “no image”).

4. Add and maintain a unique ID column

• The file must contain one column with unique values.

• This column will be used as the system ID.

5. Clean and normalize content

• Convert multi-line content into a single line.

Example:

BAD (multi-line cell)	GOOD
I worked at X • Task 1 • Task 2	I worked at X: Task 1, Task 2.

• No bullet lists or paragraph formatting.

• Remove unnecessary whitespace and formatting characters.

• Replace decorative symbols with plain text equivalents:

  • ✓ → OK

  • • → "," or "-"

  • “ ” → " (standard quotes)

6. Validate numeric fields

• Avoid scientific notation (e.g., change 1E+08 → 100000000).

7. Use correct CSV formatting

• Separator: semicolon (;) or comma (,). The portal detects it automatically.

• Encoding: UTF-8

• No blank columns

8. Missing data

• If a value is missing, you can leave the cell blank.

9. Avoid Excel auto-formatting issues

Excel may:

• Convert long numbers to scientific notation

• Remove leading zeros

• Interpret values as dates

To prevent this:

• Set ID columns to Text format before entering data

• Disable automatic formatting when possible

• After exporting to CSV, open the file in a text editor to verify correctness

Supported file formats

• Analyzed directly: TXT, CSV

• Converted through Unstructured IO: BMP, DOC, DOCX, EML, EPUB, HEIC, HTML, JPEG, MD, MSG, ODT, ORG, P7S, PDF, PNG, PPT, PPTX, RST, RTF, TIFF, TSV, XLS, XLSX

• XML: can be uploaded and opened with Preview, but not analyzed.

For unsupported formats, convert them to TXT or CSV using document processing tools before uploading.

Automatic file conversion

Smabbler Portal is integrated with Unstructured IO for automatic document processing. If you upload a file in one of these formats, the Upload file for processing in Unstructured API window opens. Enter your Unstructured API URL and key, then click Upload. The file can be up to 10 MB. The extracted text is saved as a new file named after the original, ending in -Unstructured-response.json, and you analyze it like a TXT file.

We do not store your Unstructured API URL or key.

---

Analyze

Select the files you want to analyze and click Analyze. You can analyze files with the status Uploaded or Analysis failed. TXT files are queued right away. For each CSV file, the Configure file window opens with two steps:

Step 1: Select ID column

Choose the column that contains unique record identifiers.

• The selected column must contain unique values.

• This column will be used as the system ID.

• Only one ID column can be selected.

Step 2: Select text columns

Choose one or more columns that contain the text you want to analyze.

• You can select multiple text columns.

• You can use Select all columns to quickly include all available text columns.

• The total text content per record must not exceed 2,000,000 characters.

Use column names

When enabled, values are linked to their column names, which helps when the column names give useful context.

This option is especially useful for files with many columns and rows, where column names provide important context.

For example:

Column 1 – Client name

Column 2 – Client address

Column 3 – Client phone number

Including column names helps the system better understand the structure and meaning of the data.

Starting the analysis

After selecting the ID column and text column(s), click Analyze to begin processing.

The file moves to In queue, then Analyzing, where CSV files show their progress in percent, and finally Analyzed. A notification tells you when the analysis ends, even if you have left the Files page. To stop a queued or running analysis, open the … menu next to the file name and choose Cancel analysis.

Analyzing a file uses GCUs from your balance, depending on the file size.

---

Build graph model

Once your files are analyzed, you can build your graph model.

1. Choose one or more analyzed files.

2. Click Build graph model to proceed.

3. Enter a model name, up to 120 characters, and click Create.

You are taken to the Models page. The model shows Pending create while it is being built and Inactive when it is ready to be activated.

---

File statuses

Each file has a Status column to indicate processing progress:

• Uploaded - the file is added but not analyzed yet.

• In queue - the file waits for its analysis to start.

• Analyzing - the analysis is running. CSV files show the progress in percent.

• Analyzed - the file has been processed and is ready to be used in graph model creation.

• Analysis failed - the analysis did not finish. Analyze the file again or contact support.

The Efficiency column shows the share of rows analyzed successfully.

---

File(s) actions

The toolbar above the list has Upload, Preview, Analyze and Build graph model. When your subscription is not active, these buttons are disabled. Preview opens the selected CSV or XML files in the Graph Explorer. Use the Search by tag, filename or status box to filter the list.

Open the … menu next to a file name for these actions:

• Details - shows the file's ID, name, size, creation date, last analysis date and efficiency. Download Report saves the analysis report with the numbers of successful and failed rows. The Content tab shows the text of the file.

• Remove - deletes the file after confirmation. Files that are queued or being analyzed cannot be removed.

• Cancel analysis - stops a queued or running analysis.

File's details

File's content

• Add Tags - click the plus icon in the Tags column to assign new tags to a file that currently has none.

• Edit Tags - if a file already has tags, click the pencil icon next to the tag(s) to modify or remove them. Tags will help you organize and filter files efficiently. A tag can be up to 50 characters.

• Select all - selects all files in your account, including those on other pages.

• Select visible - selects only the files currently visible on the page.

• Clear selection - deselects any selected files.

• Remove selected - deletes the selected files after confirmation. Models already built from a file keep working after you remove it.