Creating Your First API Integration

Step-by-step tutorial to wire a project endpoint into an application

Goal

Integrate one deployed endpoint into an application service with secure auth, robust request handling, and baseline observability.

Step 1: Prepare Endpoint Information

From your project, open Use this model and collect:

• Base URL
• Endpoint path
• Model name
• Endpoint type (chat, embedding, image, audio, etc.)

Step 2: Create an API key

1. Go to API Keys from the navigation.
2. Open the projects tab.
3. Create new key for the required project containing endpoint.

Step 3: Send a Test Request

curl --location 'https://YOUR_BASE_URL/v1/chat/completions' \
  --header 'Authorization: Bearer YOUR_API_KEY' \
  --header 'Content-Type: application/json' \
  --data '{
    "model": "YOUR_MODEL_NAME",
    "messages": [{"role": "user", "content": "Give me a one-line summary of Bud AI Foundry."}],
    "max_tokens": 128
  }'

Step 4: Validate Response and Move to Application Code

• Verify the response from the test request returns a 200 status code and the expected payload.
• When moving to your application code, consider enabling streaming to process the response incrementally.

Step 5: Add Retry and Error Mapping

Implement client-side handling:

• Retry 429 and 5xx with backoff.
• Surface 401/403 as credential/permission issues.
• Treat 400/422 as payload validation errors.

Step 6: Add Monitoring Hooks

Capture and store:

• Request ID (if available)
• Endpoint URL/path
• Status code
• Latency
• Token/request usage metrics

Validation Checklist

Request succeeds with valid key and payload

Invalid key returns expected auth error handling

Retry path works for simulated transient failures

Logs include endpoint, status, and latency fields

Next Steps

Proceed to advanced usage:

• Guides
• Reference
• Troubleshooting