Workflow Quick Start
A step by step guide on how to build your first Workflow in Fetch Hive.
Last updated
Was this helpful?
A step by step guide on how to build your first Workflow in Fetch Hive.
Last updated
Was this helpful?
Workflows simplify complex tasks by breaking them down into manageable steps. By connecting these steps using powerful tools like Google Search, website scraping, and document parsing, you can effortlessly build intelligent processes without writing a single line of code.
In this guide, we'll walk through creating your first workflow in Fetch Hive, configuring its capabilities, and making it ready for integration with your business applications.
To create a workflow, navigate to the Workflows section in your Fetch Hive dashboard. Here you'll need to define your start inputs - the values that will be sent to the workflow when triggering the API endpoint.
For our example, we'll create a text input called "Company name" with a corresponding key that can be referenced in subsequent steps.
Fetch Hive offers a wide variety of steps you can add to your workflow:
The Google Search step allows you to dynamically search the web using your start inputs, providing recent and relevant web results.
The AI Prompt step processes information from previous steps and generates intelligent content based on custom system and user messages.
Create images automatically using text descriptions generated in previous steps, with configurable quality settings and formats.
Deploy specialized AI agents within your workflow that can perform complex tasks based on their specific instruction sets and tools.
Retrieve recent news articles from Google News based on your search queries, ideal for timely information and updates.
Utilize Exa's semantic search capabilities to find relevant information across the web with greater contextual understanding.
Find content similar to a specific URL, perfect for research and content recommendations.
Access Microsoft's search engine results as an alternative or complement to Google Search within your workflow.
Extract content from specific websites by providing URLs, allowing you to gather and process information from across the web.
Connect to external services and applications by making API calls directly from your workflow, enabling limitless integration possibilities.
Build a custom JSON schema with all your unique data from steps in the Workflow.
Convert uploaded documents (PDFs, Word files, etc.) into searchable text that can be processed in subsequent workflow steps.
Retrieve documents from URLs and convert them to text automatically, streamlining the process of gathering document-based information.
Before deploying, you can test your workflow directly in the dashboard:
Click "Run" in the top toolbar
Enter values for your start inputs (e.g., "OpenAI" for company name)
Start the workflow run
As the workflow executes, you'll see:
Each step's progress in real-time
Execution time and credit usage
The output from each step
How data flows between steps
You can expand your workflow by combining multiple steps. For example:
Search for company information using Google Search
Generate a SWOT analysis with an AI Prompt
Create an image description with another AI Prompt
Produce a visual representation using Image Generation
All these steps can be connected and automated with a single input (company name).
Each step in your workflow offers management options:
Disable/enable steps for testing purposes
Duplicate steps to create variations
Delete steps you no longer need
Test individual steps before running the full workflow
The Activity Logs section provides complete transparency into your workflow operations:
View all workflow runs (full runs and individual step tests)
Examine inputs and outputs for each step
Monitor resource usage and costs
Analyze performance to optimize your workflows
This information helps you iterate and improve your workflows by showing exactly how data flows through each step.
Once your workflow is perfected, you can deploy it as an API endpoint:
Click "Create Endpoint" from the dropdown menu
Name your endpoint (e.g., "Example Workflow Endpoint")
A live API endpoint will be generated with a production variant
You can also create multiple variants (production, staging, etc.) to safely test changes before deploying to production.
This workflow allows you to test and refine your AI processes in the dashboard, then seamlessly integrate them with your business applications through simple API calls.