Vision (Image Analysis)

VisionAnalyzer enables the agent to analyze images using AI vision capabilities.

Supported Providers

Provider	Model	Vision Support
OpenAI	GPT-4.1 / GPT-4o	✅ Full
Google Gemini	Gemini 3.x / 2.5	✅ Full
Anthropic Claude	Claude 4.6 / 4	✅ Full
Others	—	❌ Not supported

Usage

Analyze Local Image

from instaharvest_v2.agent.vision import VisionAnalyzer
from instaharvest_v2.agent.providers import get_provider

provider = get_provider("gemini", api_key="AIza...")
vision = VisionAnalyzer(provider)

# Analyze a downloaded photo
result = vision.analyze_image(
    "photo.jpg",
    prompt="Describe this Instagram post. What's in the image?",
    language="en",
)
print(result)

Analyze Image from URL

result = vision.analyze_url(
    "https://instagram.com/..../photo.jpg",
    prompt="What products are shown in this image?",
    language="en",
)

Check Provider Support

if vision.is_supported:
    result = vision.analyze_image("photo.jpg")
else:
    print("Current provider doesn't support vision")

Parameters

analyze_image(image_path, prompt, language)

Param	Type	Default	Description
image_path	str	required	Path to local image file
prompt	str	"Describe this image in detail."	Analysis prompt
language	str	"en"	Response language

analyze_url(image_url, prompt, language)

Param	Type	Default	Description
image_url	str	required	Image URL
prompt	str	"Describe this image in detail."	Analysis prompt
language	str	"en"	Response language

Supported Formats

• JPEG (.jpg, .jpeg)
• PNG (.png)
• GIF (.gif)
• WebP (.webp)

Max image size: 5 MB