Back to .md Directory

Google Cloud

Documents steps to set up a Google Cloud RAG engine and Vertex AI Search with Dialogflow CX, including trial credits and regional agent creation.

May 2, 2026
0 downloads
4 views
ai agent llm rag eval gemini
View source

What this file does

Documents steps to set up a Google Cloud RAG engine and Vertex AI Search with Dialogflow CX, including trial credits and regional agent creation.

When to use it

  • Starting a RAG project on Google Cloud Vertex AI
  • Setting up a conversational agent with Dialogflow CX
  • Exploring Vertex AI Search pricing and trial credits
  • Configuring a custom agent in a specific region like Montreal

Assumes this stack

Google CloudVertex AIDialogflow CXRAG EnginevLLMFAISS

Google Cloud

RAG Engine

Vertex AI Search (part of Gen App Builder)

Vertex AI Search is a part of Vertex AI Agent Builder

  • https://cloud.google.com/enterprise-search?utm_source=chatgpt.com&hl=en
  • https://cloud.google.com/generative-ai-app-builder/pricing?hl=en
  • Create a project and get $1000 US of credits for 1 year - make sure to attach a billing id that does not have a pervious credit for...
  • "Trial for Gen App Builder" - Following SKUs. Vertex AI Search: Search API Request Count - Standard (services/74B1-77CF-C302/skus/BADA-EE26-7BDA) Vertex AI Search: Search API Request Count - Enterprise (services/74B1-77CF-C302/skus/93D6-7280-CF05) Vertex AI Search: Search API Request Count - LLM Add-on (services/74B1-77CF-C302/skus/151B-D7A2-B5B5) Vertex AI Search and Conversation: Data Index (services/74B1-77CF-C302/skus/BC7D-6A97-90F8) Audio interactions with Vertex AI Conversation - Voice Experience (services/FBC0-AA4A-C89A/skus/1A04-3B12-74DC) Text interactions with Vertex AI Conversation - Chat Experience (services/FBC0-AA4A-C89A/skus/90FE-6D2B-0AE1)
<img width="1855" height="809" alt="Screenshot 2026-02-05 at 14 41 42" src="https://github.com/user-attachments/assets/a8cca459-684e-42ae-b169-08ef18703f5a" /> - proceed with https://cloud.google.com/products/agent-builder?hl=en - you will see the followiing credit added to your account - Trial for Gen App Builder Available 100% $1,355.26 $1,355.26 One-time Gen_App_Builder_12b94822648ba1a2828e16bead90637469df7d0d4b62db9ffbdafe32d38cf185 Certain usage; see the terms of the promotion for details. February 5, 2026 - <img width="1598" height="125" alt="Screenshot 2026-02-05 at 14 50 38" src="https://github.com/user-attachments/assets/c5217f16-7277-432b-bb43-f7620b6fd445" />

Navigate to your gen app builder project

<img width="1855" height="552" alt="Screenshot 2026-02-05 at 14 53 54" src="https://github.com/user-attachments/assets/f06045dd-3db0-4fbe-807c-0e6fe30522c7" /> - enable dialog flow <img width="1859" height="527" alt="Screenshot 2026-02-05 at 14 54 23" src="https://github.com/user-attachments/assets/3db26e70-f5ce-400e-8614-75113a5ac48f" /> - create custom or prebuilt agent <img width="1851" height="823" alt="Screenshot 2026-02-05 at 14 56 06" src="https://github.com/user-attachments/assets/2fff24c5-8358-4e4f-8f64-beb4c21aba6a" /> - create custom in northamerica-northeast1 (Montreal) <img width="1857" height="1073" alt="Screenshot 2026-02-05 at 14 57 29" src="https://github.com/user-attachments/assets/ad3270b8-6b57-47cf-90c6-e2d814cc1203" /> - using prebuilt feedback flow - https://cloud.google.com/dialogflow/cx/docs/concept/prebuilt-component/feedback <img width="1851" height="831" alt="Screenshot 2026-02-05 at 15 00 23" src="https://github.com/user-attachments/assets/1d1e5704-c79d-4371-b4b4-66b68583aea9" /> <img width="1856" height="874" alt="Screenshot 2026-02-05 at 15 01 14" src="https://github.com/user-attachments/assets/974d6583-38b0-4b88-97bb-8e1ab0b5eeb5" />
    • todo: introduce RAG document additions and embeddings vector store

References

What's inside

2 main sections (RAG Engine, Vertex AI Search), 1 subsection, 8 screenshots, 6 links, 1 todo item

Change this for your project

  • Replace gen-app-builder-old with your own Google Cloud project ID
  • Replace northamerica-northeast1 with your desired region
  • Replace the screenshot image URLs with your own captures

Where it goes

Reference documentation for a retrieval pipeline. Keep with the ingestion or retrieval code it describes.

Worth borrowing

  • Using prebuilt Dialogflow CX components like the feedback flow to speed up agent creation
  • Linking to GitHub issues for tracking implementation progress

Related Documents