AI Assistant No API Key Needed 14 Providers Fully Offline Capable

Send an image, say a word
let AI build the label

A customer sends a screenshot of a label — or you have an old label with no source file? Hand the image to the built-in AI assistant: it first reads the fields on the image, then generates an editable template. Barcode content is decoded by the built-in recognizer instead of being guessed by the model.

Download & Try See what it does

aBiao AI Assistant: sending a label screenshot generates an editable template

Real Capabilities

What the AI assistant can do

Only what is actually shipped — no marketing fiction.

📷

Image to layout

Drop a label screenshot into the chat. The vision model reads the fields on the image (product name, spec, model, lot number, barcode content and more), lists them, then places text, barcodes and QR codes onto the canvas to produce a template you can edit directly. No more measuring and retyping field by field.

🎯

Precise barcode reproduction

This is where “image to label” usually falls apart: a large language model cannot read the byte content of a barcode, so the digits look right but scan wrong. aBiao decodes barcodes and QR codes with its built-in recognizer first, then hands the real content to the AI. Type “recognize codes” in your message to trigger it.

💬

Conversational editing

No digging through property panels. Just say “make the product name 14pt bold”, “move this barcode down 2 mm” or “take the lot number from the lot_no column”, and the assistant edits the template. Every step goes onto the undo stack.

🗃

It knows your data sources

The assistant can see the data source and columns bound to the current template, so the field mappings it suggests are real column names, not invented ones. For complex logic it can write the formula or C#/VB script and apply it straight to the component.

🆕

Zero-cost start

Pick the “DEEPSEEK Web” or “MiMo Web” channel: log in once with your own account. No API key to buy, no top-up — you use the quota you already have. No account yet? Register for free on the login page.

🏠

Runs fully offline

If data must not leave your network, point it at a local Ollama instance. The whole pipeline runs on your own machine or intranet: no per-call fees and no data leaving the plant. Pull a vision model for image-to-layout use.

Three Channels

Pick what fits, switch any time

Settings are remembered per provider.

🆕 Free web channel Zero Cost

DEEPSEEK Web · MiMo Web · No API key · No top-up

Do not want to buy an API key? Use this channel: log in once with your own account inside aBiao’s built-in browser window, then chat directly at no extra cost. Your quota is whatever that site already gives you.

Note: this channel drives the vendor’s web UI, so it cannot read images; it may also break temporarily if that site changes its front end. Switch to an API channel in that case.

No API KeyNo Top-upLog In OnceBuilt-in Browser

🔑 API channel: 14 providers

DeepSeek · Qwen · OpenAI · Gemini · Zhipu GLM · ModelScope · Volcano Engine · iFlytek · Baidu Qianfan · Tencent Hunyuan · MiMo

Image-to-layout requires this path (pick a vision model: look for v / vl / vision / omni in the name, such as glm-4.6v-flash, qwen-vl-plus or Qwen3-VL-8B). Most vendors offer free quota or new-user credits, and aBiao ships a built-in guide (“where to get free quota / where the key is”) with one line per vendor. Start with a small model to control cost.

Vision ModelsPay as you goFree-quota guideSwappable

🏠 Local / offline: Ollama

Data stays on-premises · No call fees · Air-gapped and sensitive sites

When data must not leave your network, attach a local Ollama server. Everything runs on your own machine or intranet with no per-call cost and nothing transmitted outside.

OllamaOn-premisesNo Data EgressZero Call Fee

How It Works

Three steps to get going

Nothing to read first — install and try.

Open the AI assistant

Click the AI button in the main window. The settings panel lets you change provider and model at any time.

Choose a channel and sign in

For zero cost pick “DEEPSEEK Web” and log in once via “Log in / Open window”. For image-to-layout pick an API provider and paste your key — the built-in guide tells you where to get it.

Send an image or just ask

Drag a label screenshot into the chat, or simply say what you want. To reproduce barcodes exactly, include the “recognize codes” keyword in your message.

Efficiency

With the assistant vs by hand

The difference shows up in the repetitive, error-prone steps.

Step

By hand

With aBiao AI

Customer sends only an image

—Redraw it, measure every field

✓Send the image, get an editable template

Barcode content on the image

—Copy by eye or scan then retype

✓Built-in recognizer decodes it exactly

Changing content, size, position

—Hunt through property panels

✓One sentence, one-click undo

Binding database columns

—Type column names, trial and error

✓The assistant knows your columns

Cost to get started

Must buy an API key first

✓Web channel: no key, no top-up

Send an image, say a wordlet AI build the label