Changelog

You can now create a welcome message with more than one lines. Welcome message were limited to one line.

There was missing validation and bad behavior when we had error the form was closing. Now there is new text validation and email validation that improves UX.

When you can chatbot errors we display hint to help you resolves your error without having to contact the support team.

When you scroll in the chat, a scroll button will appear. Click it to quickly return to the bottom of the chat.

Previously, users had to wait until message completion before they could scroll up. Now, users can scroll at any time.

If the disclaimer text exceeds the visible area, a scrollbar will appear, allowing you to view the entire message.

You can now display file sources in the chatbot. This will allow you to provide additional information to the user. When the user clicks on the file source, the file will be downloaded. This setting is available in the chatbot customization tab.

When embedding the chatbot in your website, you can now use your own domain. This will allow you to whitelabel OpenAssistantGPT and provide a seamless experience for your users. There is a new documentation available to help you set up your custom domain on your smart chatbot.

With the Pro plan, you can now create up to 5 custom domains. This will allow you to use multiple domains for different chatbots.

You can now integrate a button in your website to open the chatbot. This will allow you to open the chatbot in an external tab.

Get notified when your API key gets disabled. This will help you to react quickly and fix the issue.

When there is an error with the chatbot the error will be displayed in the chatbox in a red box at the buttom of the chatbox.

This feature allows users to upload multiple files at once and preview them before sending, making it easier to manage attachments.

You can now add saml/sso login on our chatbot when using our enterprise plan. You can easily configure your Azure Entra to limit who can access your chatbot

The usage page now provides a comprehensive overview of all available resources.

If you open our documentation you will see a new page called Web Search Configuration. This documenation will help you to configure your chatbot web searches correctly using the OpenAI Assistant API.

Before you had to open OpenAI console to add missing function, now OpenAssistantGPT fully manage functions for you.

Now the code interpreter tool is enabled automatically.

You will now receive email when your chatbot generates error. If needed the notifications can be disabled in your settings.

Web Search are now faster we improved the way we are using our search provider.

You can now choose the font size in your chatbot. This will allow you to have a chatbot with a custom font size.

You can now safely update your chatbot without removing you functions.

You can now configure your chatbot to be aware of the current time in the world.

We just added a new method to connect to our platform using magic link instead of having to use Github or Google.

With the new websearch feature your chatbot can now access the most recent information and the latest news about a subject.

You can now see an overview of your chatbot in the dashboard. This will allow you to see the number of messages, threads, and more.

You can now stop the chatbot when he is generating a message. You don't have to wait for the message to be generated to resend a new message.

You can now choose to add a disclaimer in your chatbot. This will allow you to inform users about the chatbot's purpose, limitations, or any other information you want to convey.

Fix issue when downloading file generated by assistant. Now the file will be downloaded correctly.

We now use our own SDK to improve the chatbot experience. This will allow us to provide better support and more features in the future.

We now support downloading the chat transcripts, users can download the chat to export it.

Now the assistant reply will not leave empty space on the right. The chatbox will be resized to fit the assistant reply.

Fix issue when assistant was searching for files but could not find them. Now tools are only enabled when needed

We now have a new chatbot SDK to help people build chatbots. The SDK is open source and can be used to build chatbots for any platform. Find out more about the SDK here https://sdk.openassistantgpt.io

Missing validation was preventing the user from updating the number of pages to crawl. This issue has been fixed and you can now update the number of pages to crawl.

We support the latest model from OpenAI, gpt-4o-mini. This model is faster and cheaper.

This feature will allow you to configure the maximum number of pages to crawl. This will prevent the crawler from crawling too many pages and using too much resources.

Exports now contain the user IP. This will allow you to see the IP of the user who sent the message in the export.

Issue was fixed and now the chat history is displaying the thread name correctly.

You can now enable the chat history in customization settings. This will allow the user to see his previous messages in the chatbot. The chat history will be displayed in the chatbot window.

You can now make chatbot fully private, only accessible by your IP. This will prevent unauthorized access to your chatbot. You can also ban specific IPs from accessing the chatbot.

Now the text input in the chatbot is a multiple line input. This will allow you to write longer messages in the chatbot. The input will grow as you write more text.

You can now choose default input style or max width input style in the chatbot. This will allow you to choose the style of the input in the chatbot.

We now use message instead of updating the thread. This will allow you to attach files to a message. The chatbot will then analyse the file and provide a response. You'll get better result

You can now set default messages for your chatbot. This will allow you to set a message that will be displayed when the chatbot is loaded. You can set this message in the chatbot IFRAME url.

Now when you need to configure OpenAI, a new modal will be displayed. This will allow you to configure OpenAI without leaving the page.

Now the text size is smaller in the chatbox. This will allow you to see more messages in the chatbox. The header is also sticky when you scroll down. This will allow you to always see the chatbox header.

We add a new CSS rule to prevent the link from overflowing the chatbox. Now the link will be displayed on multiple lines if it is too long.

We now support right to left languages in the chatbot. This will allow you to use languages like Arabic, Hebrew and more.

Better UX when uploading files in the chatbot. We now display the file name and we can also remove the file from the chat.

In the advanced settings of your chatbot, you can now configure the max_completion_tokens and max_prompt_tokens. We added validation to prevent you from entering a value lower than 256. This will prevent the chatbot from returning an error when the value is too low.

This feature is created for exporting messages from your chatbot. This feature will generate a JSON file with all the messages received by a chatbot.

Now you can copy to your clipboard the chatbot response by clicking on the copy icon in the chat.

This will prevent the chatbot from being hidden by other elements on the page.

Now in the advanced settings of your chatbot, you can configure the max_completion_tokens and max_prompt_tokens. This will allow you to control the number of tokens used in the completion and prompt.

Now we can update the chatbot customization without having an infinite loading issue.

New feature to unsubscribe from marketing emails. You can now manage your communication preferences in the settings page.

You can now try our basic and pro plan for free for 7 days. It will give you access to all the features of the plan.

We re-enabled the auto scroll in the chatbox. This will make the chatbox scroll to the bottom when a new message is received.

User in the chat can now upload files, images and more to the chatbot. The chatbot will then analyse the file and provide a response. It uses the "code_execution" function in OpenAI

When the assistant generates an image, it will now be displayed in the chat.

You will automaticly get migrated to our iframe version of the chatbot when you use the chatbotjs script.

For a better futur of the project we are removing the 3$ pricing plans. With Stripe fees, taxes and server costs we can not sustain this pricing plan. Users who already have this pricing plan will keep it and still pay 3$.

You can now update Chatbot Logo in the Chatbox, this will help you to personalize your chatbot even more.

Improve design of chatbot bubbles on user's reply.

Remove auto focus on input after message.

It is now possible to use client side prompt on your chatbot for more personalization depending on the context. Heres's how you can use it '&clientSidePrompt=You are currently talking to {User X} help him to understand the book {Book Name X}.' You can find more information in our documentation.

You can now embed the chatbox in your website using an IFrame. This will prevent CSS conflicts and provide a smoother user experience.

When you ask your chatbot to output math function and starting them with $$ it will display math function in LaTex style. Here's what your prompt should look like to display LaTex: 'Display a math function in latex and the function text must always start with $$'

We improved UI to make it look better. We also now have an integration via IFrame which means no more CSS conflicts and a smoother user experience.

You can now import an assistant that you've trained on OpenAI in our app without having to recreate everything.

We now have email notifications! You'll receive an email when you receive a new user inquiry.

Update the list of valid file extensions to match OpenAI list.

You can now see your recent user inquiry and chatbot errors in the main dashboard.

We now track chatbot errors and display them in the dashboard. We this feature you can easily see what went wrong and fix it.