Guides

Vision & Document Intelligence

Use Cortiqa models to inspect, describe, and reason over images, architectural diagrams, PDF page screenshots, and financial charts with high precision.

Capabilities

• Dense Document OCR: Extract formatted tabular data, scanned text, invoices, and receipts into structured JSON.
• Chart & Infographic Parsing: Read values, trendlines, and anomalies from business dashboards and graphs.
• UI & Wireframe to Code: Convert design mockups and whiteboard sketches into HTML/Tailwind components.
• Visual Troubleshooting: Diagnose hardware status lights, terminal error screenshots, and system error popups.

Key Use Cases

Code Example (Python)

1import base64
2from cortiqa import Cortiqa
3
4client = Cortiqa()
5
6with open("dashboard_metrics.png", "rb") as f:
7    img_b64 = base64.b64encode(f.read()).decode("utf-8")
8
9response = client.chat.completions.create(
10    model="openai/gpt-oss-120b",
11    messages=[
12        {
13            "role": "user",
14            "content": [
15                {
16                    "type": "text", 
17                    "text": "Extract all CPU, memory, and disk IO spikes shown in this chart. Format as markdown table."
18                },
19                {
20                    "type": "image_url",
21                    "image_url": {"url": f"data:image/png;base64,{img_b64}"}
22                }
23            ]
24        }
25    ]
26)
27
28print(response.choices[0].message.content)

Best Practices & Guidelines

• Native Resolution: Upload clear images without compression artifacts to preserve small numbers in tables and axis labels.
• Context Placement: Ask specific, targeted questions rather than general prompts like “tell me what this is”.
• Privacy Notice: Do not upload sensitive personal identification numbers (e.g. government credentials) without prior tokenization.