Building Agent Applications Quickly with Makers Agents
Makers Agents is a one-stop Agent development and hosting platform for developers. You can use JS or Python with mainstream frameworks such as Claude Agent SDK and OpenAI Agents SDK to write business logic. The platform automatically provides Agent runtime, model invocation, sandbox tools, session management, and end-to-end tracing capabilities. Web frontend, Agent, and business APIs can also be developed and managed within the same project. This document guides you through the complete development cycle of a Makers Agent application, covering applicable scenarios, project structure, quick start from templates, local debugging, and deployment.
What Are Makers Agents Suitable For?
Building Industry SaaS
Makers Agents is ideal for encapsulating industry knowledge, business data, and operational processes into continuously deliverable AI capabilities. Agents can not only answer questions but also invoke tools and business APIs to perform queries, analysis, generation, and process execution, helping developers build SaaS Agent products tailored to specific industries.
The following are several typical directions:
E-commerce: Build AI shopping assistants, intelligent customer service, and merchant operations assistants. Agents can understand purchase needs, recommend products, query orders and logistics, and also assist merchants in generating product copy and analyzing reviews.
Marketing and Content: Distill brand guidelines, content assets, and review standards into Agent workflows for topic planning, copy generation, multi-channel rewriting, asset analysis, and content review.
Enterprise Services: Build knowledge assistants and process assistants for customer service, sales, and operations scenarios. Connect customer, ticket, contract, or appointment systems to perform data search, information summarization, lead follow-up, and task routing.
Healthcare and Health Services: Build patient service, appointment guidance, and health management assistants. Connect medical knowledge bases, appointment, record, and follow-up systems to provide care guidance, data organization, and health knowledge Q&A.
Building a Vibe Coding Platform
Vibe Coding platforms enable end users to create, modify, and publish applications through natural language. The platform understands requirements, generates code, runs verification, and iterates continuously, ultimately delivering applications that remain accessible over the long term. Products such as Vercel v0 and Lovable all fall into this category.
EdgeOne Makers integrates core capabilities such as production, deployment, compliance detection, access, and multi-tenant isolation into a single platform. This enables platform providers to establish the complete Vibe Coding chain of "conversation generation → real-time preview → application launch → multi-tenant isolation" without building and maintaining the underlying infrastructure themselves. For the detailed solution, refer to Building a Vibe Coding Platform.
Platform Advantages
Out of the Box: All tools, memory, observability, and models required for Agent execution take effect automatically after deployment, with no installation or integration needed.
Flexible and Open: Choose any framework (Claude / OpenAI / LangGraph / CrewAI) with no language restrictions (JS / Python).
Seamless Integration: Web and Agent share the same project, with unified management of accounts, deployment, monitoring, and domains.
Understanding How Makers Agent Works
Makers Agents adopts a development model of "dual-runtime collaboration + unified platform capability injection". The Web frontend, Agent, and business APIs can be placed in the same project, sharing deployment, domain, and observability. The platform selects the appropriate runtime for different workloads based on the directory where the code resides. For more information, refer to Agents Overview.
my-agent/├── agents/│ └── customer-service/│ ├── index.ts # POST /customer-service│ └── stop.ts # POST /customer-service/stop├── cloud-functions/│ └── conversations/│ └── list.ts # GET /conversations/list├── app/ or src/ # Frontend application├── edgeone.json # Build configuration and Agent framework declaration└── .env.example # Declares variable names only and does not store real keys.
Quick Start
Template selection
Makers provides two types of templates:
Template Type | Applicable Scenarios |
Framework base template | Provides a minimal runnable starter template based on various Agent frameworks, suitable for extending business logic from scratch. |
Business scenario template | Includes basic features and common business interactions, and is suitable for quick launch after branding is modified and business processes are integrated. |
If you need to build and deeply customize your business from scratch, we recommend choosing a framework-based template. If you want to quickly validate your product, you can directly use the business scenario template that best matches your requirements. For more available templates, see Agent Template List.
Creating a Project via the CLI
You can also install EdgeOne Cli by using the following command and quickly create a basic template based on the OpenAI Agents SDK.
npm install -g edgeoneedgeone makers create --template openai-agents-starter-node
Creating or Migrating Existing Agent Projects with Skills
Makers provides Skills for Agent development and migration. For the detailed installation process, see Skills. After the corresponding Skill is loaded in development tools such as Cursor and CodeBuddy, AI can complete Agent project creation and migration according to Makers platform specifications.
You can directly enter prompts similar to the following into the AI:
Build a conversational Agent on EdgeOne Makers with the Claude Agent SDK.
Transform this LangGraph Agent project into the EdgeOne Makers platform format.
Local Debugging and Deployment
Starting Local Development and Debugging
If you need to use environment variables already configured in the Makers console, run
edgeone makers link first to link an existing project.Start local debugging:
edgeone makers dev
You can access the Agent page at http://localhost:8088/.
Access the local observability tracing page at http://localhost:8088/agent-metrics.
Deploying a Project
If the project is already linked to a Git repository, pushing code triggers Makers' CI build and deployment. You can also deploy directly through the CLI:
# Deploy to Productionedgeone makers deploy -n <project name>
After a successful deployment, click the link in the Console to access the specific build information and the deployed URL.
