Generator
Generate text using input prompt and selected LLM model.
Overview
This workflow implements a chatbot that enables users to upload an image and ask questions about its content. Using a combination of Optical Character Recognition (OCR) and visual recognition technologies, the chatbot analyzes the image and provides accurate, context-sensitive answers. This automation is highly valuable for scaling tasks where users need to extract information from images or interact with visual data conversationally.
Step-by-Step Flow
Chat Initialization
- When the chat session is opened, the workflow triggers a welcome message using the Message Widget.
- The message introduces users to the chatbot’s capabilities, explaining that they can upload images and ask questions about the content.
User Input Handling
- Users can interact with the chatbot by:
- Typing a question about an image.
- Uploading an image file.
- The Chat Input node captures both the question (text message) and the uploaded image (file input).
- Users can interact with the chatbot by:
Image and Question Processing
- The Generator node receives:
- The uploaded image (for OCR/visual recognition).
- The user’s question (as context for the large language model).
- The generator analyzes the image, extracts information (e.g., text via OCR or visual features), and formulates a relevant answer to the question.
- The Generator node receives:
Response Delivery
- The answer generated by the model is routed to a Chat Output node, which displays the response to the user in the chat interface.
- If an image was uploaded, it can also be displayed in the chat for reference.
Workflow Structure
Here’s a simplified structure of the workflow:
| Step | Node Type | Function |
|---|---|---|
| Chat opened | ChatOpenedTrigger | Triggers the welcome message |
| Display welcome message | MessageWidget | Shows introduction and instructions |
| Show message to user | ChatOutput | Presents the welcome message in chat |
| User inputs question / uploads image | ChatInput | Collects user text and image file |
| Process image & question | Generator | Performs OCR/visual recognition, answers query |
| Display generated answer (and image) | ChatOutput | Shows the answer (and possibly image) to user |
Benefits and Use Cases
- Automation & Scalability: This workflow automates the process of extracting information from images, enabling rapid and consistent answers to visual questions without human intervention.
- Versatility: Useful for customer support, educational tools, document analysis, and any scenario where users need to query or understand images.
- Enhanced User Experience: Provides a conversational interface, making it easy and intuitive for users to interact with complex image analysis tools.
- Seamless Integration: The modular node-based design allows for future expansion or integration of more advanced recognition models.
Example Use Cases
- Document Digitization: Users upload pictures of documents and ask for summaries or specific details.
- Product Support: Customers send images of products and inquire about specifications or issues.
- Educational Tools: Students upload diagrams or charts and ask explanatory questions.
By automating visual question answering with this workflow, organizations can make powerful image analysis tools accessible to a broad audience, reduce manual effort, and deliver faster, smarter responses at scale.
