Skip to main content
This guide walks you through creating a new endpoint from scratch.

​
Before You Start

Think about:
  • What will this endpoint do? — Summarize text? Classify content? Generate responses?
  • What inputs does it need? — Text, numbers, options, files?
  • What outputs should it return? — Structured JSON? Simple text?

​
Creating the Endpoint

1

Navigate to Endpoints

Click Endpoints in the sidebar to open the endpoints list.
2

Click Create

Click the Create Endpoint button in the top-right corner.
3

Fill in Basic Details

Complete the form:
FieldDescriptionExample
NameHuman-readable nameEmail Subject Generator
PathAPI URL path (must be unique)/api/v1/generate-subject
CategoryOrganizational groupEmail
DescriptionWhat this endpoint doesGenerates compelling email subject lines from email body content
4

Set Visibility

Choose who can see this endpoint:
  • Public — All team members can view and use
  • Internal — Only visible to you and admins
5

Save

Click Create to save your endpoint.

​
Path Requirements

Your endpoint path must follow these rules:
All paths begin with /api/v1/ for consistency and future versioning.✅ /api/v1/summarize
❌ /summarize
No two endpoints can have the same path within your tenant.
Paths should be lowercase with hyphens separating words.✅ /api/v1/generate-subject
❌ /api/v1/GenerateSubject
Good paths describe what the endpoint does.✅ /api/v1/classify-sentiment
❌ /api/v1/do-the-thing

​
After Creation

Once created, you’ll land on the endpoint detail page. From here you should:
  1. Define Input Schema — Specify what data your endpoint accepts
  2. Define Output Schema (optional) — Specify the response structure
  3. Create a Prompt — Write the LLM instruction
  4. Test — Verify everything works
  5. Set Default — Promote and set the default prompt
An endpoint without a default prompt will return an error when called. Make sure to create and set a default prompt before using it in production.

​
Example: Building a Sentiment Analyzer

Let’s create a complete endpoint for sentiment analysis:

​
Step 1: Basic Details

FieldValue
NameSentiment Analyzer
Path/api/v1/analyze-sentiment
CategoryAnalysis
DescriptionAnalyzes text and returns sentiment (positive, negative, neutral) with confidence score

​
Step 2: Input Schema

Add one required field:
FieldTypeRequiredDescription
textStringYesThe text to analyze

​
Step 3: Output Schema

Define the expected response:
FieldTypeDescription
sentimentStringOne of: positive, negative, neutral
confidenceNumberConfidence score between 0 and 1
explanationStringBrief explanation of the classification

​
Step 4: Create Prompt

Analyze the sentiment of the following text and classify it as positive, negative, or neutral.

Text to analyze:
{{ inputs.text }}

Respond with a JSON object containing:
- sentiment: "positive", "negative", or "neutral"
- confidence: a number between 0 and 1 indicating your confidence
- explanation: a brief explanation of why you chose this classification

Be accurate and objective in your analysis.

​
Duplicating Endpoints

Need a similar endpoint? You can duplicate an existing one:
  1. Find the endpoint in your list
  2. Click the ⋯ menu
  3. Select Duplicate
  4. Modify the name and path
  5. Adjust schema and prompts as needed
Duplicating is great for creating variations (e.g., summarize-short and summarize-detailed) that share similar structures.

​
Deleting Endpoints

Deleting an endpoint is permanent and will break any applications using it. Make sure nothing depends on the endpoint before deleting.
To delete:
  1. Go to the endpoint’s Settings tab
  2. Scroll to the Danger Zone
  3. Click Delete Endpoint
  4. Confirm by typing the endpoint name

​
Next Steps

Define Input Schema

Specify what data your endpoint accepts

Define Output Schema

Specify what data your endpoint returns