This post is my submission for DEV Education Track: Build Multi-Agent Systems with ADK.
What I Built
I built an AI Blog Writer which accepts a title and style/prose of writing and generates textual content in that style for the blog.
Cloud Run Embed
Your Agents
There are 4 Agents + 1 Orchestrator to coordinate between the agents
i) Researcher
This receives the user input topic/title. It produces a structured brief covering key concepts, trends. It doesn't write in prose and just gathers foundation for other agents' working.
ii) Outliner
This agent receives the topic and research brief from previous agent. It is a content strategist which determines the structure : titles, sections, conclusion. It writes only the skeleton
iii) Writer
This agent receives the data from previous agent and produces long form prose sentences and writes the complete draft.
iv) Editor
This agent receives the draft and validates it. Hook Strength, Clarity, Flow, Grammar, Tonal Consistency. It outputs final polished blog post.
Key Learnings
The capability of the agents to stick to their specific role was insightful. The technical depth of the blog and transformation of the prose was also a reflection of the capability.
One of the challenging part was to get all of it together and deploy due to IAM issues on Cloud. But was able to resolve all, and successfully deploy.
Top comments (0)