Documentation - Files

OpenAssistantGPT Files

The upload file option allow you to upload a table of question and answer that trainer wants your chatbot to learn how to respond. The system would provide your chatbot with more knowlege and skills in such way that tainer can upload a releveant file that his chatbot needs to learn by answering the question or by reading the table. And when chatbot is retrained with that file it will enhance chatbot performance such as replies related to the content of the uploaded file.

To upload a file on our platform, follow these steps:

  • Open: https://openassistantgpt.io/dashboard/files/upload
  • Upload a file.

Note that upload types that are not files are currently not supported by the system. Allowed upload extensions are: 'c', 'cpp', 'csv', 'docx', 'html', 'java', 'json', 'md', 'pdf', 'php', 'pptx', 'py', 'rb', 'tex', 'txt', 'css', 'jpeg', 'jpg', 'js', 'gif', 'png', 'tar', 'ts', 'xml', 'zip'. Upload files with unsupported extensions will be prohibited.

Client-Side File Uploads

OpenAssistantGPT uses client-side file uploads powered by Vercel Blob storage for both the dashboard file uploads and chat attachments. This implementation offers several advantages:

  1. Larger File Support: Files up to 500MB can be uploaded directly from the browser to Vercel's storage, bypassing the 4.5MB Vercel Serverless Function size limit.

  2. Improved Performance: Files are uploaded directly from the client browser to the storage service, reducing server load and improving upload speeds.

  3. More File Types: Our system supports a wide range of file types for chat attachments, including documents, images, audio, video, and code files.

When using the chat interface, users can attach files by clicking the paperclip icon. Files are uploaded directly to storage and appear as attachments in the chat. For larger files, a progress indicator will show the upload status.

Technical Implementation

Behind the scenes, we use Vercel Blob's client-side upload functionality, which creates a secure, direct connection between the user's browser and Vercel's storage infrastructure. This approach:

  • Reduces server load by offloading upload processing
  • Allows for larger file uploads than would be possible through API routes
  • Maintains security through token-based authentication
  • Ensures files are properly associated with the correct user and conversation
CrawlersChatbots