[FREE] ChatGPT with all text/voice models, Computer Vision and Dall-e 2 and 3 - Openai API

Extensions
,
#61

Hi, does the extension support receiving of text streams?

1 Like
#62

I will add it in the next update.

1 Like
#63

Thanks! Do you know of any extensions that support WebSockets and headers?
Is your extension free or a sponsor version?

#64

Try this

#65

Hi, thanks! I got it. It looks like what I was looking for :slight_smile:

1 Like
#66

Thanks, I hope so ,
And I I am waiting for your feedback :sparkles:

#67

Hi, do you have a demo.aia? the one in the download can’t load in kodular as a project.
I’m focused in the gpt stream part.
image

Br Teddy

#68

Try with MIT app inventor the aia file originally for AI2 not kodular

#69

It’s loading perfectly in App Inventor - Thx.

1 Like
#70

17. UploadTrainingFile (NEW)

  • Description: Uploads a JSONL file for use in fine-tuning.

  • Parameters:

    • apiKey (Text): Your OpenAI API key.
    • filePath (Text): Path to the JSONL file to upload.

  • Returns: Triggers FileUploaded event with the file ID or ReportError if an error occurs.

  • How it works: This method sends a multipart/form-data HTTP POST request to OpenAI’s API to upload the specified JSONL file for fine-tuning purposes.

18. StartFineTuningJob (NEW)

  • Description: Starts a fine-tuning job using an uploaded JSONL file.

  • Parameters:

    • apiKey (Text): Your OpenAI API key.
    • trainingFileId (Text): The ID of the uploaded JSONL file.
    • model (Text): The base model to fine-tune (e.g., gpt-4o-mini or gpt-3.5-turbo).
    • suffix (Text): Optional suffix to append to the fine-tuned model’s name.

  • Returns: Triggers FineTuningJobStarted event with job details or ReportError if an error occurs.

  • How it works: This method sends an HTTP POST request to OpenAI’s fine-tuning endpoint, creating a fine-tuning job using the provided file and model.

3 Likes
#71

Thanks, Ahemed, I have pm’d you.

1 Like
#72

Creating GPT JSONL files

Click here to access my website that generates the files in an intuitive way.

#73

NEW - Version 9 (GPT Embedding)

image

#74

News

The extension was factored into categories and recompiled with Fast Cli and ProGuard.

image
image1366×768 100 KB

1 Like
#75

I´m trying to use Vision with an image from a component but I´m geting error all the time, only working with URL images, any ideas?

Captura de pantalla 2024-12-18 144947
Captura de pantalla 2024-12-18 144947668×527 93.6 KB

Screenshot_20241218_135520
Screenshot_20241218_1355201080×471 94.8 KB

#76

Unfortunately, it is a problem with the MIT App Inventor base system itself, which does not allow working with base 64 images.

1 Like
#77

How did I solve this problem by sending to an Image2URL conversion server?

#78

do you know any free api service to do that?

#79

I designed a solution with Imgur and Imbb server: [FREE] Convert images to URL with Imgur API

2 Likes
#80

thank you so much! I will check it out