Integration with AI Features

How to Use Vectors in Diaflow Vector Node

Prerequisites: You have vectors uploaded and an automation flow created.

Search Data from Vectors:

1. Add Diaflow Vector node to your flow

2. Set Action to “Search data from Vectors”

3. Choose data source:

• All data from vectors: Searches entire knowledge base

• From vector groups: Select specific groups

• From vector files: Select specific files

1. Configure Input field with search query or dynamic reference

2. Connect to LLM node(s) to analysis the data which queried at Diaflow Vector node

3. Setup output and execute the flow

Add Data to Vectors:

1. Set Action to “Add data to Vectors”

2. Choose Destination:

• Root: Adds to main directory

• Group: Select target group

1. Configure Input with file path or URL

2. Set Advanced Configurations:

• Chunking Method, Chunk Size, and Chunk Overlap

1. Connect LLM model(s)

2. Setup output and execute the flow

Expected Result: Data is processed and becomes available for search operations.

UI Enhancements:

• Selected files/groups display count (e.g., “Selected 6 files”)

• View Files/Groups button shows detailed selection list

How to Integrate with Chat Simple Mode

Prerequisites: You have vectors uploaded and want to create a knowledge-based chatbot.

Steps:

1. Navigate to Chat, Simple Mode

2. In Knowledge Source section:

• Select Vector data sources from Resource - Vector

• Choose scope: Root, Group, or Files

1. Configure AI Engine (provider, model)

2. Set up personality and prompts

3. Check by asking some questions.

4. Publish your chatbot if you want everyone to be able to access it.

Data Integration:

• Vector context appears as {{vec-0.output}} in prompts

• Follow-up questions consider both conversation and vector context

Expected Result: Chatbot responses include relevant information from your knowledge base.

How to Use Vectors in Page Generation

Prerequisites: You want to generate documents (blogs, contracts) using your knowledge base.

Steps:

1. Navigate to Page on the sidebar, then generate document type

2. Complete initial input step

3. Click Setting icon

4. Knowledge source screen appears, select data you want to use

5. Setup AI Engine

6. Use a prompt to search the knowledge base you selected

7. AI generates content incorporating vector knowledge

Source Integration:

• Documents: Summarized paragraphs used as context

• Articles: Main body text provides supporting ideas

• URLs: Text extracted without HTML/scripts

• Groups: Aggregates most relevant vectors

Expected Result: Generated content reflects knowledge from selected vector sources.