Vision in (Un)Perplexed Spready — Analyze Images Directly in Your Spreadsheet

Introduction

(Un)Perplexed Spready is a spreadsheet application that combines the power of traditional spreadsheets with advanced artificial intelligence capabilities. Version 1.2.X represents the biggest leap in the application's history, bringing a range of features that elevate data work to an entirely new level.

This version introduces four major improvement areas: Web Search, Vision & Document Processing, SQL Query, and Advanced Statistics.
Best of all, SQL Query and Advanced Statistics functions are completely free! If you are looking for a free statistical software, you are on the right page!

Spreadsheet applications traditionally work with numbers and text. But in the real world, many data come in visual form — charts, tables, receipts, documents, product photos, logos, diagrams...

(Un)Perplexed Spready v1.2.X introduces Vision functionality — the ability to bring images into AI context directly from formulas in your spreadsheet. Source of images can be local folder or an URL.

How It Works

Vision functionality uses Vision LLM models (such as qwen3-vl, llama3.2-vision, etc.) that can "see" and analyze images. You simply specify the path to an image (local or URL) in a formula, and the AI analyzes the content and returns the result to the cell.

At the moment of writing this article, we can recommend usage of qwen3-vl:235b-cloud model available in Ollama Cloud.

Basic syntax:

=ASK_LOCAL1("IMAGE:/path/to/image.png", "instruction")
=ASK_LOCAL1("IMAGE:https://example.com/image.jpg", "instruction")

Supported Image Formats

Format	Extensions	Notes
JPEG	.jpg, .jpeg	Most common format
PNG	.png	Lossless, transparency
WebP	.webp	Modern format, good compression
GIF	.gif	Animated images supported
BMP	.bmp	Windows bitmap

Usage Methods

1. Local Images

Images stored on your computer:

=ASK_LOCAL1("IMAGE:/home/user/Documents/receipt.png", "Extract total amount from this receipt")
=ASK_LOCAL1("IMAGE:C:\Users\User\Pictures\chart.png", "Analyze this chart and describe trends")
=ASK_LOCAL1("IMAGE:./data/table.png", "Convert this table to structured format")

2. Online Images (HTTP/HTTPS URLs)

Images from the internet:

=ASK_LOCAL1("IMAGE:https://example.com/product.jpg", "Describe this product")
=ASK_LOCAL1("IMAGE:https://charts.example.com/sales.png", "Analyze this sales chart")

3. Combining with Other Data

=ASK_LOCAL2(A2, "IMAGE:/path/to/product.png", "Compare the description in A2 with the product image")
=ASK_LOCAL3(A2, B2, "IMAGE:https://example.com/chart.png", "Analyze chart in context of data in A2 and B2")

Practical Applications

Receipt and Document Analysis

=ASK_LOCAL1("IMAGE:/scan/receipt_001.jpg", "Extract: date, store, total amount")
=ASK_LOCAL1("IMAGE:/invoices/invoice.pdf_page1.png", "Extract invoice number, customer, and amount")

Reading Tables and Charts

=ASK_LOCAL1("IMAGE:/reports/chart.png", "Read values from this chart and list them")
=ASK_LOCAL1("IMAGE:/data/table.png", "Convert this table to CSV format")

Product Categorization

=ASK_LOCAL1("IMAGE:/products/item_001.jpg", "Categorize this product: category, subcategory, color")
=ASK_LOCAL1("IMAGE:https://shop.com/product_123.jpg", "Describe product: name, features, price if visible")

Document Analysis

=ASK_LOCAL1("IMAGE:/docs/id_card.jpg", "Extract: name, surname, ID number (mask last 3 digits)")
=ASK_LOCAL1("IMAGE:/contracts/contract_page1.png", "Summarize main points of this contract")

QC and Inspection

=ASK_LOCAL1("IMAGE:/qc/sample_001.jpg", "Assess quality: are there visible defects?")
=ASK_LOCAL1("IMAGE:/production/widget_123.jpg", "Check: is this product correct?")

Cross-Sheet Range with Images

You can combine images with range data from other sheets:

=ASK_LOCAL2("IMAGE:/products/batch_001.jpg", "RANGE:Products!A2:A50", "Compare image with product list and identify which one is in the image")

Configuration

Required Models

Vision functionality requires a Vision LLM model that supports image processing.

Recommended models:

qwen3-vl — Excellent for OCR and document analysis
llama3.2-vision — Good balance of speed and quality
llava — Open-source vision model
minicpm-v — Fast and efficient
kimi-k2.5
qwen3.5:397b

Installing models (Ollama):

ollama pull qwen3-vl
ollama pull llama3.2-vision

Settings > AskLocal Settings

API Endpoint: URL of your Ollama server (usually http://localhost:11434)
Model: Vision model (e.g., qwen3-vl)
API Key: If required

Limitations

Image Size

Recommended size: up to 2048x2048 pixels
Larger images are automatically scaled
Very large images may cause timeout

Number of Images

One image per formula
For multiple images, use multiple formulas

Real-World Examples

Data Entry Automation

You have a folder with scanned receipts:

Cell A1: =ASK_LOCAL1("IMAGE:/receipts/001.jpg", "Date")
Cell B1: =ASK_LOCAL1("IMAGE:/receipts/001.jpg", "Store")
Cell C1: =ASK_LOCAL1("IMAGE:/receipts/001.jpg", "Total amount")

Tips for Best Results

Use clear instructions — The more precise your question, the better the answer
Be specific — Specify exactly what you want to extract
Control the format — Request JSON, CSV, or tabular format
Use quality images — Clear images give better results
Choose the right model — qwen3-vl is excellent for documents, llama3.2-vision for general analysis

Error Messages

Error	Meaning
FILE NOT FOUND	Image does not exist at the specified path
INVALID IMAGE	Image format is not supported
APIERR	Model not available or does not support vision
TIMEOUT	Image too large or server overloaded

Privacy and Security

Local models: Images stay on your computer
Remote models: Images are sent to remote server

Conclusion

Vision functionality in (Un)Perplexed Spready opens the door to an entirely new way of working with data. You are no longer limited to text and numbers — you can now analyze visual content directly in your spreadsheet.

Whether you're processing receipts, reading tables from photos, categorizing products, or analyzing charts — Vision formulas give you the power of AI vision at your fingertips.

Vision is a PREMIUM feature. Requires Vision LLM model (local or cloud).

Home

(Un)Perplexed Spready v1.2.X — Analyze Images Directly in Your Spreadsheet