Dify Integration

AgentQL integrates with Dify, allowing you to flexibly build AI applications.

Dify helps to create production-ready AI solutions, from agents to complex AI workflows with low-code and user-friendly interfaces.

Set up AgentQL with Dify

  1. Sign up for a free Dify account, from there you can jump right in.
  2. Go to the Dify dashboard.
  3. On the top of the dashboard, open the Tools tab
  4. Search for "AgentQL".
  5. Click on the AgentQL tool.
  6. Click on "To authorize", provide your AgentQL API key to use the AgentQL tool, and click "Save". Generate your AgentQL API key here.

Usage

The AgentQL tool is available for use in both workflows and agents.

Using AgentQL in an Agent

For agentic usage, you'll add AgentQL's Extract Web Data tool to your agent app like so:

  1. Go to the Dify dashboard.
  2. Click "Create from Blank" or "Create from Template" to create a new app and choose an "agent". A workspace will open up.
  3. Under Tools, click "+ Add".
  4. Search for "AgentQL"
  5. Select the "Extract Web Data" tool under AgentQL.

Next you'll prompt the agent to extract data from a web page with either an AgentQL query or a Natural Language prompt.

Using AgentQL in a Workflow

For workflow usage, you'll add AgentQL's Extract Web Data tool to your pipeline like so:

  1. Go to the Dify dashboard.
  2. Click "Create from Blank" or "Create from Template" to create a new app and choose a chatflow or workflow. A workspace will open up.
  3. On the bottom left, click the "Add Block" button.
  4. Under Tools, search for "AgentQL"
  5. Select the "Extract Web Data" tool under AgentQL.

Next you'll configure input variables in the tool's UI, and run the pipeline to extract web page data.

Support

If you have any questions about using AgentQL with Dify, you can get involved in the Dify community to ask questions.

Examples

  • Price Deal Finder: In the chat input, name a product that you would want to find and compare prices of. You can start off with the example below:
terminal
Nintendo Switch - OLed Model - w/ White Joy-Con
  • Research Assistant: In the chat input, ask any question you have about any research topic. You can start off with the example below:
terminal
What are some interesting facts about black holes to research on?
  • News Aggregator: In the chat input, provide a prompt of what you want to scrape and URL link for the website to scrape. You can start off with the example below:
terminal
Scrape all job postings from the following page: https://www.ycombinator.com/jobs
Include columns for: Job Title | Company | Location | Job URL | Employment Type (Full-time, Part-time, Contract, etc.) | Remote Eligibility (Yes/No)