Integration Guide

Detailed implementation for registering external pipeline

Registration Process

This section will introduce to you on how you can create your own pipeline service

Step 1: Implement Required Endpoints on Your Service

1. GET /verify - Returns service metadata

2. POST /responses - OpenAI-compatible chat endpoint

See External Pipeline Requirements for detailed specifications.

Step 2: Deploy Your Service

Deploy your service with:

• Publicly accessible URL (HTTPS recommended for production)

• Valid API key configured

• Endpoints tested and working

Step 3: Register with GLChat

Call GLChat's registration endpoint:

curl -X POST "https://glchat.example.com/register-external-pipeline" \
  -H "Authorization: Bearer YOUR_GLCHAT_TOKEN" \
  -H "Content-Type: application/json" \
  -H "X-Tenant-ID: your-organization" \
  -d '{
    "url": "https://your-service.com",
    "api_key": "your-service-api-key"
  }'

Or you can add external pipeline via Admin Dashboard

Step 4: Wait for Verification

GLChat will:

1. Call your /verify endpoint

2. Validate response format

3. Create pipeline infrastructure, including prompt catalog, lmrp catalog, and default preset

4. Publish configuration event

Step 5: Create Chatbot

1. Go to GLChat Admin Dashboard

2. Create new chatbot application

3. Select your registered external pipeline as pipeline preset

4. Start chatting!