Create Your First Agent

The ChainOpera AI Developer Platform provides multiple ways to create AI agents, from simple prompt-driven assistants to fully customized agents powered by workflows and frameworks. This flexibility allows you to start quickly and scale to advanced, production-ready use cases.

Accessing the Agent Creation Page

  1. Log in to your account.

  2. Navigate to the sidebar and click “Create Agent”.

  3. Choose one of the four creation modes:

  • Prompt – Fast and simple prompt-driven agents

  • Workflow – Visual, multi-step logic with drag-and-drop nodes

  • Third-Party API – Integrate agents or services from external platforms

  • Agent Framework – Full code control for advanced developers

Prompt-based Agent

Best for: Quickly creating single-purpose AI agents using natural language.

Overview:

A Prompt Agent allows you to define the agent’s behavior with a single or multi-line instruction. You can enhance it further with Knowledge Bases and MCP Servers for added context and capabilities.

Steps to Create:

  1. Select Prompt in the creation menu.

  2. Enter your Agent Name and Description.

  3. Write your prompt instructions to define the agent’s role and output style.

  4. (Optional) Attach a Knowledge Base for personalized context.

  5. (Optional) Enable MCP Services to add external tools or APIs.

  6. Test your agent in the playground and publish when ready.

Example Prompt:

You are a friendly financial assistant.
Analyze the user’s expenses and provide suggestions for saving money.
Respond in a concise and actionable style.

Workflow-based Agent

Best for: Agents that require multi-step logic, data processing, or tool chaining.

Overview:

The Workflow Builder allows you to design complex agent logic visually. You can combine multiple steps like prompt processing, conditional branching, and API calls without writing full code.

Steps to Create:

  1. Select Workflow from the creation page.

  2. Start from scratch or choose a Workflow Template to speed up setup.

  3. Drag and drop nodes to build logic, such as:

    • Prompt Node – LLM processing

    • API Node – External service calls

    • Condition Node – Branching logic

    • Tool/MCP Node – Execute modular compute or sub-agents

  4. Connect nodes to define the workflow path.

  5. Test the workflow in the playground, verify outputs, and save.

  6. Publish your workflow-based agent to make it available to users.

Use Case Example:

  • A Document Q&A Agent can:

    ① Accept a file from the user → ② Extract key content → ③ Search in KB → ④ Return summarized answers.

Third-Party API Agent

Best for: Reusing existing AI agents or integrating real-time data from external services.

Overview:

This mode allows you to bring agents from other platforms or connect external APIs to your ChainOpera agent.

Steps to Create:

  1. Select Third-Party API in the creation menu.

  2. Enter the API endpoint and authentication details (API Key, OAuth, etc.).

  3. Map request inputs and response outputs to your agent.

  4. (Optional) Add fallback logic using the workflow editor.

  5. Test and publish the API-driven agent.

Example Use Case:

  • Connect a live weather API to provide real-time forecasts.

  • Integrate a CRM API to fetch customer data and create a support agent.

Agent Framework

Best for: Developers who want full control, custom logic, or multi-agent orchestration.

Overview:

With Agent Framework, you can upload your own Python-based agents or use frameworks like AutoGen or AgentOpera. This method is ideal for creating highly customized and performance-intensive agents.

Steps to Create:

  1. Select Agent Framework from the creation menu.

  2. Upload your agent code or connect a Git repository.

  3. Configure the runtime environment, including:

    • Entry script

    • Dependencies

    • Environment variables

  4. Test locally or in the platform’s sandbox environment.

  5. Deploy to Agent Servers for persistent execution.

Future Support:

  • LangChain and CrewAI will be supported for broader multi-agent and orchestration capabilities.

Tips for Selecting a Creation Method

  • Prompt → Quick experiments or single-purpose bots.

  • Workflow → Multi-step processes or automation tasks.

  • Third-Party API → Leverage existing services or integrate external data.

  • Agent Framework → Full customization, complex logic, or long-running tasks.

PreviousIntroductionNextAgent Management

Last updated

Was this helpful?