agentusecasesAll 965 use cases
Software & DevOps

Orchestrate a multi-phase RAG migration

A walkthrough of using Gemini CLI to coordinate a complex, multi-phase cloud migration project.

Done withGemini CLI

What they did
The author ran Gemini CLI with the Conductor extension. Markdown files defined the product, tech stack and workflow. For each track the agent wrote a spec and plan, which he approved before implementation. It then worked test-first, wrote Terraform, and ran checkpoints at the end of each phase. He ran it in a sandbox in Yolo mode with a notifier script, and returned when notified.
How it went
The project took several days of AI-assisted work. It used about 19M input, 66M cached input and 400k output tokens, costing around $30 on Vertex AI. The source gives no failure details.
Worth knowing
Cached input tokens dominated usage, and the author says costs vary. Rules like running terraform plan before apply and the coverage target belong in workflow.md.

Try it yourself with Claude Code

I'm migrating [system, e.g. my search and RAG pipeline] from [current setup] to [target cloud service]. Read the codebase in this folder and break the migration into phases with clear checkpoints, risks, and rollback steps. Finish with a written plan in MIGRATION.md; don't change any code or run any cloud commands until I approve each phase.

Read the original ↗

Source: DEV · Apr 28, 2026

Five of these in your inbox every morning

The best things people got an AI agent to do, each with the prompt to try it.

More like this