Create Universal OpenAI-Compatible API Endpoints for Multiple AI Workflows

This n8n workflow provides a flexible and universal API endpoint that can be used to integrate with various AI services, making them compatible with the OpenAI API specification. It acts as a proxy, receiving requests, processing them, and then forwarding them to the appropriate AI service based on the incoming request's content.

What it does

This workflow simplifies the process of interacting with different AI models by:

1. Listening for incoming API requests: It uses a Webhook to receive HTTP POST requests, acting as the universal endpoint.
2. Extracting AI model information: It processes the incoming request body to identify the target AI model or service.
3. Routing to the correct AI workflow: An 'If' node evaluates the extracted model information and directs the request to a specific n8n sub-workflow designed for that AI model.
4. Executing AI logic: For a specific AI model (demonstrated with an "AI Agent" and "OpenAI Chat Model"), it processes the chat messages, potentially using memory for conversational context.
5. Making external API calls (if needed): The "HTTP Request" node is available for making calls to actual AI service APIs.
6. Responding to the caller: It aggregates the results and sends a formatted response back to the original caller via the Webhook.

Prerequisites/Requirements

• n8n Instance: A running n8n instance to host and execute the workflow.
• OpenAI API Key (or compatible AI service credentials): If you intend to use the OpenAI Chat Model or similar services, you will need the appropriate API keys/credentials configured in n8n.
• Understanding of OpenAI API Specification: Familiarity with the OpenAI chat completion API structure will be beneficial for configuring incoming requests and expected responses.

Setup/Usage

1. Import the Workflow:
   • Copy the provided JSON code.
   • In your n8n instance, go to "Workflows" and click "New".
   • Click the three dots next to the workflow name and select "Import from JSON".
   • Paste the JSON and click "Import".
2. Configure the Webhook:
   • The "Webhook" node (ID: 47) is your universal API endpoint. Activate the workflow to get its URL. This URL is what external applications will call.
   • Ensure the "HTTP Method" is set to POST.
3. Configure AI Service Credentials:
   • For the "OpenAI Chat Model" (ID: 1153) or any other AI service you integrate, you will need to set up the corresponding credentials in n8n.
   • Click on the "OpenAI Chat Model" node, and select or create a new "OpenAI API" credential.
4. Customize AI Routing Logic:
   • The "If" node (ID: 20) is crucial for routing. You will need to modify its conditions to detect which AI model/service the incoming request is for. This typically involves checking a field in the incoming JSON body (e.g., json.model).
   • For each branch of the "If" node, you would connect to a separate sub-workflow or a series of nodes designed to interact with a specific AI service.
5. Implement AI Service Workflows:
   • The current workflow includes an example path using an "AI Agent" (ID: 1119), "Simple Memory" (ID: 1163), and "OpenAI Chat Model" (ID: 1153). This path demonstrates a basic conversational AI setup.
   • For other AI services, you would add new nodes (e.g., "HTTP Request" nodes to call their APIs, "Code" nodes for data transformation) on different branches of the "If" node.
6. Test the Endpoint:
   • Once configured, activate the workflow.
   • Send a POST request to the Webhook URL with a JSON body that includes the necessary information for your AI model (e.g., {"model": "gpt-3.5-turbo", "messages": [{"role": "user", "content": "Hello!"}]}).
   • Monitor the workflow execution in n8n to ensure it routes and processes requests correctly.