Prompt Quick Start
A step by step guide on how to build your first prompt in Fetch Hive.
Last updated
A step by step guide on how to build your first prompt in Fetch Hive.
Last updated
Prompts play a vital role in generating awesome output for the Fetch Hive platform. In this guide, we will go through a simple example on how to set up your first prompt and link it to an endpoint, ready for your business to utilize.
Let's dive in...
Log in to your Fetch Hive account
Navigate to "Prompts" in the left hand sidebar
Click the "Create Prompt" button in the top right corner
Set a name for your prompt
Additionally, you can associate a Category to easily organize your prompt library
Choose the AI model (Anthropic or OpenAI)
Select the specific model (e.g., GPT-4, Claude)
Set the temperature (controls randomness of output)
Set the max tokens (limits the length of the response)
Add a system message (e.g., "I am a helpful assistant")
Write your main prompt (e.g., "Create a 100 word summary on {topic}")
Use placeholders like {{topic}} for dynamic inputs
Click the "Inputs" button
Specify input variables (e.g., "topic")
Add test values for these inputs
Click "Run" to test your prompt
Review the generated response
Go to the drop-down menu in the top left corner
Select "Create Endpoint"
Alternatively, you can update an existing endpoint by click "Update Endpoint"
Give your endpoint a name (e.g., "SummTopic")
Review the endpoint summary and parameters
Set a fallback model (e.g., Claude if OpenAI is unavailable)
Add any collaboration notes if needed
Click the "Code" button in the top right corner
Ensure you include necessary input values when calling the endpoint
Run the API call in your console or application
Verify that the response matches your expectations
By following these steps, you can create a custom prompt in Fetch Hive and link it to an API endpoint for easy integration into your projects. Remember to iterate and refine your prompt as needed to achieve the desired output.