- Services Installation Guide
- Gen AI Document Analyzer Service Installation
- Application Installation
- Configure Gen AI service with OBRH
11.1.2 Configure Gen AI service with OBRH
This topic describes the systematic instructions to configure the Gen AI service with OBRH.
Calls to the LLM can be routed via OBRH. Follow the below steps to configure OBRH to use the Gen AI Service.
Note:
Before configuring the OBRH, make sure that all the common dependencies are installed except the LLM dependencies.Note:
For detailed information on the OBRH configuration, refer to Routing Hub Configuration User Guide.- On Home screen, click Core Maintenance. Under Core Maintenance, click Routing Hub. Under Routing Hub, click Service Consumers.
- Create a new service consumer with default name GENAI.
- Click on GENAI service consumer and add a new service
provider with the following details:
Table 11-1 Add Service Provider
Field Value Product Name COHERE Type EXTERNAL Version 0.1 Headers Add the below headers as Name - Value: - accept - application/json
- content-type - application/json
Service Update the below values under service Type OTHERS Name Add the below services: - CHAT_ENDPOINT
- EMBED_ENDPOINT
Http Method POST Endpoint Add the below endpoints for the respective services: - For CHAT_ENDPOINT- v1/chat
- For EMBED_ENDPOINT - v1/embed
Service Header Add the service header for both the services: - Name - Authorization
- Value - Bearer $body.token
- Click the COHERE service provider and add the
implementation parameters.
Table 11-2 Add Implementation
Field Value Name COHERE_Default Description Default Implementation Type DEFAULT Default Toggle Enabled Scheme https Host api.cohere.com Port 0 - Click Consumer Services and add the following service
IDs.
- COHERE_CHAT_ENDPOINT
- COHERE_EMBED_ENDPOINT
- Click the consumer service IDs and add the transformation, and routing
details.
- COHERE_CHAT_ENDPOINT
Table 11-3 Add Transformation
Field Value Name Chat Active Toggle Enabled Product Processor COHERE 0.1 Implementation COHERE_Default Service CHAT_ENDPOINT - v1/chat Request Transformation Update the below values under request transformation Body Type RAW Template Type JSLT Template {"model": .model,"message": .message, "temperature": .temperature, "chat_history": .chat_history,"prompt_truncation": "AUTO","stream": false,"connectors": []} Table 11-4 Add Route
Field Value Name chat Start/Stop START Auto Shutdown Toggle Enabled Default/Custom Rule Default Rule Transformations Click Add and select the transformation created.
- COHERE_EMBED_ENDPOINT
Table 11-5 Add Transformation
Field Value Name embed Active Toggle Enabled Product Processor COHERE 0.1 Implementation COHERE_Default Service EMBED_ENDPOINT - v1/embed Request Transformation Update the below values under request transformation Body Type RAW Template Type JSLT Template {"texts":.texts,"model": .model,"input_type": .input_type} Table 11-6 Add Route
Field Value Name route1 Start/Stop START Auto Shutdown Toggle Enabled Default/Custom Rule Default Rule Transformations Click Add and select the transformation created.
- COHERE_CHAT_ENDPOINT
Parent topic: Application Installation