tcxGPT

OpenAI Responses API client addon for TrussC.

Port of ofxResponsesAPI for TrussC.

Features

• Asynchronous HTTP requests via background thread
• OpenAI Responses API (chat completions)
• Token usage tracking
• Configurable model, temperature, and system instructions
• JSON response format support

Dependencies

• tcxCurl (included in TrussC addons)

Setup

1. Add tcxGPT and tcxCurl to your project's addons.make:

   tcxCurl
   tcxGPT
   

2. Run projectGenerator to update your project.

3. Create bin/data/secrets.json with your OpenAI API key:

   {
       "openai_api_key": "sk-proj-..."
   }

Usage

#include <TrussC.h>
#include <tcxGPT.h>
using namespace std;
using namespace tc;
using namespace tcx;

GPT gpt;

void setup() {
    gpt.setup("YOUR_API_KEY");
    gpt.setModel("gpt-4o-mini");
    gpt.setInstructions("You are a helpful assistant.");
}

void update() {
    if (gpt.hasNewResponse()) {
        auto response = gpt.getNextResponse();
        logNotice() << response.text;
        logNotice() << "Tokens: " << response.totalTokens;
    }
}

void keyPressed(int key) {
    if (key == ' ') {
        gpt.sendMessage("Hello, how are you?");
    }
}

Advanced: Custom Request

GPT::Request req;
req.input = "Translate to Japanese: Hello world";
req.model = "gpt-4o";
req.temperature = 0.3;
gpt.sendRequest(req);

API Reference

GPT

Method	Description
setup(apiKey)	Initialize with API key
sendMessage(input)	Send a simple message
sendRequest(request)	Send a custom request
hasNewResponse()	Check for pending responses
getNextResponse()	Get next response from queue
setModel(model)	Set model (default: gpt-4o-mini)
setTemperature(temp)	Set temperature (0.0-2.0)
setInstructions(text)	Set system instructions
setMaxOutputTokens(n)	Set max output tokens
setResponseFormat(json)	Set structured output format

GPT::Response

Field	Type	Description
text	string	Response text
id	string	Response ID
status	ResponseStatus	Completed/Failed/Incomplete
errorCode	ErrorCode	Success or error type
errorMessage	string	Error details
totalTokens	int	Total tokens used
inputTokens	int	Input tokens
outputTokens	int	Output tokens

Example

See example-chat/ for a complete working example.

License

MIT