Image Testing — Multiple Scenarios

Test various image generation and manipulation use cases
API Documentation →

Validate checks API access only. If Generate Image returns 401, enable Images API (DALL·E) for your key in the OpenAI dashboard.

Scenario 1: Text → Image

Generate an image from a text prompt

Use Case: Generate images from text descriptions using OpenAI's image generation API.

Features:

  • Support for prompt templates with variable substitution
  • Multiple image sizes (512x512, 1024x1024, 2048x2048)
  • Automatic text generation to pair with the image
  • Custom filename support for generated images

How it works: Enter a text prompt describing the image you want to generate. The system calls OpenAI's image generation API and optionally generates a short descriptive text to accompany the image.

Idle

Result

Generated image and metadata

Generate Video (Gemini Veo)

Create a short video from a text prompt

Use Case: Generate videos from text using Google Gemini (Veo). Output format matches the image API: trusys.output with video_url and text.

API key: Use the Gemini API key field above (from Google AI Studio), or set GEMINI_API_KEY / GOOGLE_API_KEY in .env.

Idle

Video Result

Generated video and metadata

Scenario 2: Edit an Image

Modify an existing image with a prompt

Use Case: Edit or modify existing images by providing a text prompt describing the desired changes.

Features:

  • Upload PNG or JPEG images, or provide an image URL
  • Optional mask file for precise editing control
  • Automatic generation of edit description text
  • Preserves original image while creating edited version

How it works: Upload an image and provide a prompt describing the changes you want. Optionally upload a mask (PNG) to specify which areas should be edited. The system uses OpenAI's image editing API to create a modified version.

Idle

Edit Result

Edited image and metadata

Scenario 3: Text + Image from Prompt

Generate both text and image from a single prompt

Use Case: Generate both descriptive text and an accompanying image from a single prompt, useful for content creation.

Features:

  • Simultaneous text and image generation
  • Separate model selection for text and image generation
  • Coherent pairing of text and image content

How it works: Provide a single prompt, and the system generates both a descriptive paragraph (using Chat Completions) and a matching image (using Image Generation API) that complement each other.

Idle

Text + Image Result

Generated text and image

Add PDF Document

Extract text and page images from a PDF

Use Case: Process PDF documents by extracting text and rendering pages as images for RAG ingestion.

Features:

  • Automatic text extraction from PDF
  • Page rendering to PNG images (up to 5 pages)
  • Automatic captioning of rendered pages
  • Multimodal embeddings combining text and visual content

How it works: Upload a PDF file. The system extracts text using pdf-parse, renders pages as images using pdfjs-dist and canvas, captions the page images, and creates embeddings for retrieval.

Note: Requires server dependencies: pdf-parse, pdfjs-dist, and canvas. If unavailable, only text extraction is performed.

Idle
Note: Requires server to have pdf-parse, pdfjs-dist, and canvas installed. If rendering is unavailable, only text is used.

Query RAG Store

Search and retrieve information from stored documents

Use Case: Query the RAG store to retrieve relevant documents and images based on semantic similarity.

Features:

  • Semantic search using cosine similarity
  • Retrieves top-K most relevant documents
  • Generates contextual answers using retrieved content
  • Returns representative images from retrieved documents

How it works: Enter a query prompt. The system embeds your query, finds the most similar documents using cosine similarity, uses the context to generate an answer, and returns relevant images from the top documents.

Idle

RAG Result

Retrieved information and context