Tutorials > AI API Calls

Head over to the folder/server/api/ai any file with.ts extension is is your API endpoint! ✨ 🤖 We've included a custom retry logic to help you navigate the unpredictable nature of AI APIs. 🔄 This ensures that your system can handle AI API outages smoothly – all for your convenience. 👊

In your new api-route.ts file, you can set up an example AI API call like this and enjoy the benefits of our retry logic working behind the scenes. ✨

/app/api/ai/gpt.ts

      try {
        // Custom retry logic considering the outages in AI APIs
        axiosRetry(axios, {
          retryDelay: (retryCount) => {
            return retryCount * 1;
          },
          retries: 20,
        });

        const res = await axios.post(url, body, options);

        const answer = res.data.choices[0].message.content;
        const usage = res?.data?.usage;

        console.log(">>> " + answer);
        console.log(
          "TOKENS USED: " +
            usage?.total_tokens +
            " (prompt: " +
            usage?.prompt_tokens +
            " / response: " +
            usage?.completion_tokens +
            ")"
        );
        console.log("
");

        return answer;
      } catch (e) {
        console.error("ChatGPT Error: " + e?.response?.status, e?.response?.data);
        return null;
      }
    