Imagine, if you are selling the bot to a consultant or trainer with some exclusive knowledge. We are training the bot with his specialist knowledge and he or she should have the ability to charge the bot-user for getting one-on-one custom answers for his or her problems/questions. For this there should be messages counter and billing for a set of messages using STRIPE pay Wall.
Enable preloaded questions/buttons in the chat widget to streamline the UX (e.g., 'How do I book?', 'How much?', 'Directions', 'Locations', 'Tell me about your products', etc.)
Another great idea to consider is having TEMPLATES for the following use-cases
Onboarding (Customers as well as Employees Two separate templates)
Training.
Lead Nurturing
Internal Data Analysis
Automated Sales Assistant
You can also think of some more use cases. Thanks 👍
In "Settings", let us add different OpenAI API's to different Chatbots. Us who runs multiple companies, which have nothing to do with etch other, need them to integrate diffrently (and billing on two different companies).
When we detect a bot is hallucinating or provide erroneous information, we should be able to interrupt the bot's answer to save some tokens in case of long answers.
Need some easier to understand tutorials that are less technical by providing some specific Use Case scenarios. Especially important is to provide guidance on when it is best to use an embedded thread versus the full chatbot embed. Look at Pabbly to see how they do their use case tutorials.
The videos that don't have audio / subtitles are VERY hard to follow / understand.
It would be nice to be able to customize the elements of the bot (header, conversation bubbles, input field) with your own design values (font color, background color etc)
Should really fix trhe White Label (retune name in source code, Tire level showing in source code, "lifetime" showing in source code, and Class ID, etc).
+++++++++++++++++
Also, a Client Dashboard, where they can operate their chatbot and now see all the extras (Retune logo, textes, roadmap, recipes, etc.)
e.g. limit each user to 10 per day. We can setup a payment link to Stripe/Paypal etc. Once paid, they receive a promo code to key into the chatbot to unlock more messages.
Also think about having the ability to insert videos, Ads ( Single and Carousel) inside the chat automatically at certain intervals.
If you can provide these your app will become more attractive and compelling. Thanks 🙏
For example, when you start the chat it says "How can I help" 2 buttons come up, one says "Support" and the Other says "Sales", If you choose sales it loads the Sales persona or if you choose support it loads the Tech Support persona.
Example:
1. Customer clicks email link - https://retune.so/.../?name=John&problem=fever
2. Re:tune injects the tags to the initial prompt:
"You are a chat-doctor and your patient is {{name}} and he has {{problem}}. Please advice him with his situation."
By @Mahtileka
It would be a significant Pro when users could order bots on the front end within the embedding in your own site. So we can sell the bots by a webshop. Like WooCommerce
I want to be able to choose Llama 2 models as an alternative to GPT. This could be included in the plans („only“ maintenance and server costs, I guess) or sold at a much lower price point than even GPT 3.5 Turbo.
For B2B enterprise and consulting projects, need the ability to clone for specific clients or use cases. Allows for testing before pushing live. Simple playground will not be sufficient for more complex use cases. Clone workspaces will be an essential feature for this platform to scale with developers who do B2B projects or working with enterprise clients who need to do extensive testing before pushing a release live.
Enabling chained requests allows for steps in answering user questions. think of user asks a question, first answer is the embedded data, second answer is found in a database (API request), third answer is found through a google search API etc.
Ability to allow users to interact with a chatbot that stores only a memory of this individual's interactions. This would create a unique profile for each user.
Would be able to identify what the user is looking at and offer complimentary products or offer additional information about the product the viewer is looking and chatting about. Could also be used to check for competitor prices across multiple sites.
We kindly request the implementation of a new feature that would enable website administrators to receive a chat transcript notification via email whenever a user engages with the chatbot on our website. This functionality would provide us with a convenient and efficient way to stay informed about user interactions in real time.
Benefits of the Proposed Feature:
Real-time Insights: The ability to receive chat transcript notifications via email would allow us to gain real-time insights into user interactions, enabling us to better understand their needs, preferences, and pain points.
Improved Responsiveness: With email notifications, we can promptly respond to user queries or concerns, enhancing the overall user experience and demonstrating a high level of customer service.
Data Collection: By receiving chat transcripts directly in our email, we can easily collect and organize valuable data for analysis, helping us refine our chatbot's responses and improve its effectiveness over time.
Outside of sales and customer service bots, I would love the ability to share some other types of very helpful chatbots with others who would greatly benefit from them, but the GPT-4 model is too expensive - though truly outstanding compared with GPT-3.5 and other models. I can share some fantastic chatbots with many people if they could input their own OpenAI API key and pay for their own token usage. Can you provide this option, please?
I would like to be able to send content to my bot via email to enable automatic processing. Both the content of the mail plain text/html and its attachment pdf, xls, docx.... should be able to be taken over.
Experience seamless communication with a chat interface that responds using voice, enabling API connectivity to third-party services for a natural and interactive conversation.
connect the N8N.io platform through a webhook, this platform is opensource, n8n allows you to build flexible workflows focused on deep data integration
Add more filter options in the thread section (like last week, last month,..) and then add an export button.
The threads could be exported to csv. So i can analyse them, let ChatGPT cluster and sum the most important questions
To understand what is going on in conversations is currently hard work - we have to individually select a thread, and read it, Imagine if we could query threads using a Bot front end (for one bot or across all bots) to ask questions like:
1) "Show me the top ten questions being asked"
Most common questions and answers. We could easily see which answers need more training.
2) "Show me the most frequent 20 keywords for all questions being asked."
Most common keywords
3) Filter by specific date: e.g. "What is the most popular question today / last week?"
- Ability to change color or theme (from purple) of the entire admin panel
- Ability to change tag-line (this can be used to identify retune)
- Ability to add logo
Let's say I've trained my bot on 20 product pages. If a visitor asks the bot to recommend a product, the bot should provide a clickable link for each appropriate result. (I believe this may also cover the requested 'citations' feature?)
Make is a popular alternative to Zapier, but it offer more flexibility & some more advanced features. It is the 2nd most popular in the marketplace for automation tools
Does it work with UCHAT?
I'm using ChatBase and installed a mini app that makes the integration a lot easier.
I'm part of a community in Brazil with more than 2136 automation professionals, many are on UCHAT for simplicity and one of the DEVs created a mini APP that helps a lot.
I believe that if you have this integration, many will abandon CHATBASE, because RE:TUNE defeats the resources developed so far, imagine with these updates?
Would love to get the bot to ask specific questions of the user such as name, email, phone, custom question 1, custom question 2, custom question 3 which we can send via webhook or simply store as a user/chat session
When inviting clients, re:tune should send a customisable email with a custom sender domain to clients providing a link to set their own password. Or at least, we're being able to set a password during client setup within a workspace, instead of sending out an email everytime someone wants to login
Current Behavior:
In the current system, there isn’t a clear way to sort conversation logs by date and time.
Desired Behavior:
I would like to request a feature that allows users to sort conversation logs by date and time. Users should have the option to sort in both ascending and descending order. This will enable us to quickly find specific conversations or observe patterns over time.
Benefits of the Feature:
Improved efficiency: Users can quickly locate specific conversations, saving time and effort.
Enhanced usability: The feature will make the platform more user-friendly and intuitive.
Better insights: Observing patterns over time can provide valuable insights for improving communication strategies or identifying issues.
Suggestions for Implementation:
A sortable column for date and time in the conversation logs view.
Options to sort in both ascending and descending order.
The sorting preference should be saved for each user so they don’t need to adjust it every time they access the logs.
Thank you for considering this feature request.
Sometimes we end up mixing different chatbots in the same workspace, would be nice to be able to migrate the entire chatbot to another workspace without losing any of its content or training. The only way to do this right now is to recreate the chatbot which takes a lot of time. Additionally while migrating to a different workspace, be able to keep the trainings and settings and connect it to a different OpenAI key.
I propose to consider developing the Retune ecosystem with an additional marketplace for selling ready-made bot templates.
1. Retune can take a commission for transactions, and this is an additional, possibly very significant income
2. Retune users who know about promt-engineering will be able to sell their templates. And users without promt-engineering experience will be able to quickly deploy their projects.
A lot of people here have problems with customization. At the same time as people who know how to customize bots, do not use their full potential.
The train feature is important and needs more attention. Here are my suggestions for an upgrade:
- hide option for deleted items
- rename option for uploaded files, texts, links
- modify option for uploaded text documents, texts
- view+modify option for text from scraped websites
- view+modify option for text taken from youtube videos
- option to group uploaded items to categories
- multiple selection of uploaded items for deleting, categorizing, moving to another category,
- time stamp for uploaded or/and modified items
- refresh button for website links to update the data if needed
- comment/description field for each uploaded item
- token counter for each uploaded item
and possible further suggestions from the roadmap comments
For my specific use-case, I need access to data in real-time. I can see the web client is capable of streaming the results so I would like access to that API
Possibly an option to view stats for each chat and also stats for each embedding (# of tokens/characters, etc., # of URLs indexed, # of docs indexed, etc.
A drag and drop UI would be honestly a mic drop for Retune. I have to say I am very impressed with the workflow and the ease of us of how Voiceflow's UI works to create a bot. Please check them out here if you havent seen it. voiceflow.com if you haven't seen it you should.
Outbound Webhooks when:
- Conversation started (inc. user IP, location, channel: e.g. Twilio, Telegram, Facebook, WhatsApp, etc)
- Conversation Responses
- Conversation ended
- Data captured
Inbound Webhooks to:
- Trigger a rescrape of a pdf, website, csv, google doc, etc
Outbound Webhooks should allow:
- Authorisation (OAuth, Basic, None)
- Headers (Any amount of custom headers)
- Body (Should be customisable, using outputs as variables)
At the account level, be able to set a default custom domain for all workspaces (and still have the ability to customize at the workspace level, if desired)
Please enable the possibility to generate a new workspace and set up the domain (I understand this requires an extra step by the client to set up their DNS). Also, enable an endpoint to set up chat, define the key, the model, the prompt, the restrictions, and the embedding data.
I want to be able to sell my openai tokens (for example, I might have credits available) or buy from others (for example, I might live in a country where openai is not available or I might need more tokens than approved in my account quota)
Imagine, if you are selling the bot to a consultant or trainer with some exclusive knowledge. We are training the bot with his specialist knowledge and he or she should have the ability to charge the bot-user for getting one-on-one custom answers for his or her problems/questions. For this there should be messages counter and billing for a set of messages using STRIPE pay Wall.
Is there a way to customize any error messages from the chatbot so that it says to contact us rather than Retune? My users won't know who/what Retune is.
I run a membership site and would like to crawl the forums and resources in the site. All this info can only be accessed when you are logged into the site.
Feature to give database credentials and maybe with a json schema file to enable the chatbot to search in a db. With this feature : The ability to set conditionals (aka field "client_id", search only in data with client_id 1.
(Maybe i'm dreaming lol)
As we continue to improve our customer support chatbot, it is crucial to monitor its interactions for quality assurance, troubleshooting, and ongoing training. This feature request outlines the need for a chatbot logging system that records all conversations the bot has with users, complete with time stamps for each exchange.
This feature is essential for monitoring, improving, and maintaining the quality of our chatbot interactions, and thereby, our customer service as a whole.
Usage Cases:
1) As a Customer Support Manager, I want to review chatbot conversations, so that I can evaluate the chatbot's performance and identify areas for improvement.
2) As a Data Analyst, I want to access time-stamped logs, so that I can analyze customer interaction patterns over time.
3) As a Developer, I want to check error logs, so that I can debug issues effectively... etc.
Ability to restrict interacting with bot, via Password or even Username/Email and Password.
Reasons/use cases:
- Confidential Knowledge
- Monetizable Knowledge
- Prevent abuse
Privacy laws in Europe are very strict. Using a chatbot and sending private data to a server outside europe could cause a break of these laws. User need to be sure to use a privacy protected chatbot.
To stop users putting anything or fake emails in to the lead capture form please consider email verification. Emails could be verified in under 0.5 seconds via the Reoon Email Verifier API.
So rather than just the Chatbot interface can there be just an input box interface (Smart Input box) that one can put on a homepage for example as an input box on a website with a big title on top saying "Ask me anything about (a particular topic )", the training is still the same just a different interface, This could be especially useful for Custom Domain and White-labeling Tier
I'm quite active on Quro, and the Questions and Answers there can form as a database of my responses. Would be great if Re:Tune and pull the Q&A from my quora profile as an embedding. And to refresh that on a constant basis so that all the answers I provide automatically becomes part of the embedding.
This is very important for non-english users. Make sure we can customize the warning/conformation text. The text is in English, but not all (end)-users are English.. I'm talking about this text (including the buttons): 'Are you absolutely sure?'
'This will permanently delete all messages in this conversation' 'Cancel' 'Reset'.
Add an option for users to flag off-topic or inaccurate responses. Site admins can review the list of flagged responses for context improvement and bug monitoring.
Instead of doing a video at a time it would be great to be able to put in a channel URL and have it train on the videos on that channel. Or atleast pull in a list of the videos from the channel similar to the "From a webpage" option.
This would make building a chatbot for a Youtube creator a breeze and be super useful to their audience.
The ability to call external APIs to send POST, GET requests, etc. Example: Making an API call to update an AirTable or to get flight info, weather, or any information provided by APIs.
Example of API capability from a competitor: https://www.youtube.com/watch?v=uk1O3zIlHuA
As a user I want to be able to toggle on and off the ability for the chatbot to suggest follow up questions to keep the user engaged and avoid confusion.
- Toggle follow up questions on or off for a chatbot
- Ability to turn on only for initial greeting to get the conversation started or for all responses
- Fixed examples - I provide a set of example follow up questions that the agent can choose from
- Few-shot generative - I provide a set of examples that the bot uses to generate new questions based on the current conversation
Every time i type my answer in the chat window and the bot answers, the window looses the focus, so that i have to click again into the field before i can type the next anser. That blocks the conversational flow feeling
Like in other competitor (f.ai in appsumo for example), would be very useful to deploy a chrome extension connected to a specific chatbot for my client. So the user can login and connect with the reserved bot (trained with business user information) and assist him writing an answer on google maps reviews or an email etc.
As website owner, I i would like to be able to monitor to which time my chatbot talked to website visitors.
I would like to be able to see topics splitted into calendar entries.
Is it possible to add a data source like Google Docs and as and when the doc is updated the bot will auto-train itself? I am not a tech-savvy person, but this may mean the bot will be connected to the data source via API or some other way and at a certain interval it will grab information for that source (similar to Zapier)
Would be stunning if we had an advanced no-code bot designer, just as Voiceflow or Botpress allow to create their complex bots and prototypes with collaboration features and integrations: https://www.voiceflow.com/conversation-design
This could go in many directions down to billing for usage system, but to keep it simple and to start.
How about a simple calculation that keeps a usage history for each bot?
it could be worked out on the character count and AI token price, giving resellers a rough idea of how much each client is spending each month.
This whole idea would be eliminated if you are going to implement individual APIs per chatbot as we can then use the Open Ai api Usage chart.
Users should be able to upload their dataset and then probably can fine-tune the data from the dashboard itself to get more precise answers. generating and formating the data would be a huge help as well
As a user I want to define the structure of a conversation, and multiple agents with access to different data sources to be responsible for each stage of the conversation, so that I can clearly define the flow, state, and objectives of the conversation.
The text is extracted from the image, and the bot is trained on the text extracted. As OpenAI continues to roll out Dall-E 3’s image processing to their ChatGPT API, continue to explore image processing capabilities.
I want a feature where I can calculate how much a user has consumed ( total number of text ) so that I can charge the user accordingly. After the user has totally used up all the text then the chat bot would show a message to buy more ..... from plan
I think its a no-brainer that at some point if you are creating embeddings that are shared data for different clients that you don't want to have to retrain the exact same embeddings in different workspaces. This could save a significant amount of time and tokens if this work does not have to be repeated.
1. Email contents to a chatbot ; so each chatbot needs to have its own unique email address where PDF, TXT files can be set and also URLs can be sent which can be parsed
2. RSS feed can be ingested; regularly; not sure if this is working currently via the URL functionality.
Mistral 7B (beta) is a free model supported by OpenRouter.ai. Adding Mistral 7B to the list of options for re:tune to create chatbots is a good idea because of its high performance and versatility. The model excels in tasks involving code and English language, making it ideal for a wide range of chatbot functionalities. Its ease of fine-tuning also allows for quick adaptation to specific user needs, providing a more personalized and efficient chat experience.
It could be important that bot could know real time and date, many customers want to know if the shop is open or close with a simple question like "Are you open" and Bot says we are open from 9am to 6pm, thats good, but sometimes, it says NO, and we really are open, but it take a time from somewhere and say NO,
As a developer I want to embed the chat widget as a full-width chat UI in my application so that I can provide users with a dedicated chat page and saved conversations like ChatGPT.
Example:
- I have a SaaS product with a sidebar and main container
- I want to add a menu section in sidebar named Chatbot with nested Conversations (retrieve all threads from retune API and display thread name)
- User clicks on Conversation
- Chat UI opens in main container, full width and height
- Chat displays full conversation history up to last message
To prevent misusage of the chatGPT API by excessive use or spamming a possibility to add an API token per chatbot or client to prevent heavy usage of our openai tokens.
Define current date/time by client-side date or by date+timezone setting in retune app as backup if client-side date is inaccessible. Use Case: User asks "What events are coming up later this month/day?" Current Problem: The response is void of current date/time and therefore either doesn't provide an answer or is inaccurate. So, in retune app, add a date and time selector with timezone so the system itself can pass along the date/time with the prompt (if unable to determine client-side date/time).
Crisp is a powerful customer messaging platform that can greatly enhance our website’s customer support and engagement capabilities. By integrating Crisp, we can provide real-time assistance to our website visitors, answer their queries promptly, and ultimately improve their overall experience.
I believe that integrating Crisp with our website would be a valuable addition to our customer service strategy. It would allow us to streamline our communication processes, increase customer satisfaction, and ultimately drive better results for our business.
Can you copy these guys? It's called HelpGent
https://wpwax.com/helpgent/
HelpGent allows you to have asynchronous video messaging to communicate with anyone, anytime, without needing to be available at the same time.
Get the power to communicate with your audiences/customers/clients/desired ones asynchronously with the most personalized communication system using video messaging, voice messaging, text messaging, and screen recording.
Without the possibility of customization, the widget does not make sense for business purposes :( Please make it possible to use your styles and assets, as well as insert the widget through a custom domain
Without the possibility of customization, the widget does not make sense for business purposes :( Please make it possible to use your styles and assets, as well as insert the widget through a custom domain
Imagine if a bot would carry on a conversation, ask and collect data. And then ask the question “I have collected all the necessary data, do you want me to send an application?”. The client presses yes and the application is sent to telegram, whatsapp, email, sms, or simply added to the table (at least, the table is enough in the basic version)
I'd like the chatbot to be able to respond with audio such that people could speak to it with mic speech to text and have the bot reply with text to speech for a full audio conversation
Ability to create a link for an embedding that can be shared and edited with your clients or team. Basically, it's a shareable page with a text box for the embedding. It can be edited anytime and will update the chatbot once saved/published. This embedding turns the dynamic text data from the textbox into static text, allowing the Al to parse it whenever necessary. Please also allow us to white-label these share links.
Similar to WhatsApp, I would love to have a Slack integration. Depending on the channel, we have different bots. For example, for the sales team or for marketing, etc.
I would love to be able to download all the conversations that my chatbot is having with clients.
it helps me undertand their need and it helps me see if the chatbot is answering properly.
I know I can see them online, one thread at a time from my account, but I want to have them in a pdf or csv for easier analysis.
I've been told that client access to bots is coming, but I can't find it on roadmap. it's a must.
I want my clients, for whom I create the chatbot and install on their website, to be able to
- see the chat history for their chatbot
- download the chat history
- add some embeddings to their chatbots
- make corrections to erroneous answers given by the bot.
They need to be able to log in and have acces to their bot only, without having to go throuhg my account.
Would be awesome to have a native WooCommerce integration, as a Real Time Knowledge source. (Without going through an Webhook).
Be able to configure and select via Toggle switches, which fields and custom fields to learn from. Particularly, when custom fields beyond Price, Description, Stock etc are available.
For GDPR purposes, user should be able to give consent of giving conversations access to the data processor.
by opening the chatbot, it should have first a UI that checks the consent button before proceeding to the chat.
If user does not consent, the data processor should not have access to the conversation.
What if the Workspaces behaved like folders? Enabling us to move a bot from "in development" workspace to "published workspace" or a certain client's workspace.
Hi Re:tune Team,
Able to add an extra column for the roadmap? E.g. between "Idea" and "In a month" *add "Future development". The purpose is to bring over the "idea" to Future development, knowing what the team will add in the future will bring confidence to many Sumo
As a user I want the chatbot to provide its sources as a list of links in the message so that I can verify the source of information and explore it in more detail. If the source is a website, I want to click on a link to go to that website in a new tab. If the source is a document (PDF), I want to click on a link and download the document.
Creating a new workspace and inviting a collaborator is the way to proceed to create a chatbot for a client and invite them to collaborate. But we always have to start from scratch with every new workspace.
Please, in creating a new workspace, give us the option to duplicate an existing one so we don't have to re-enter our logo, embeddings, chatbot rules, etc, every time.
obviously we may want to change some things after, but having the option to duplicate an existing workspace would be a big time saver.
Allow to set up chatbot as an internal support bot to help answer emails. Give the ability to instruct reply of email copied text. i.e reply to this email saying...
Capability to connect a team of agents/chatbot of research agents capable of web browsing, email communication, shared memory utilization, and much more.
This will be a super feature.
We require an intermediary role that sits between the 'owner' and 'member' roles. This new role should have the capability to add new embedding data and members without requiring approval from the workspace owner. Ideally, this could be a 'manager' role with configurable settings for each workspace. For instance, the owner could set member and embedding limits when creating the workspace.
Please let us change the white label appearance like color of panel, menu and login screen with our own logo. Option to add custom links in the menu in iframe (plus new tab) would be nice to put in our own tutorial videos, support chat, community or what ever we want. ;)
It would be great to have the options to view the documents or references like Afforai does, if the user toggles that mode on, basically build it out to use this if they want to
https://appsumo.com/products/afforai/?utm_source=google&utm_medium=cpc&utm_campaign=GS-Buy-Prospect-Dedicated-Select-Manual-WorldWide-Experiment-Branded-Product-Deals&utm_adset=160169783452&utm_content=678250636071&utm_term=afforai&loc_i=&loc_p=9058788&network=g&dvc=c&utm_placement=))&gclid=Cj0KCQjw4vKpBhCZARIsAOKHoWTeXKzS6kEVIsVI4WwvPZfxsTjaIk2taKmL74KiJ322bDHa8qdd6A4aAlEaEALw_wcB
I am looking into buying their product as well, however, if you have on the road map very soon I will remain here, I dont see why Retune can not have this as a toggle feature for some of the chats to let users view referenced material etc. just like Afforai, as many others have said, this would be so powerful Retune & Afforai had a baby.
Please upvote this, as this would be a powerful feature.
Add a reporting feature that enables the tracking of chat metrics (e.g., Duration, number of questions and number of replies until exit, total number of chats, time of day)
Allow users to create chat threads. A single continuous chat history (like WA and TG) would be ok. Multiple threads would be best. Use case: Multiple therapy/coaching sessions. Multiple threads will make for shorter chat history per thread hence the bot will be able to take the whole session into context.
curiousthing.io chatbots can answer phones and book meetings and send follow up texts using chat gpt, among other services, i feel like this would take retune to the moon, thank me later
A Feature rich Chatbot Builder integrated with various parameters and can be customized based on difference purposes.
It's just like "creating a form" to build a chat bot in order to maximize its use.
It would be awesome if Retune could generate a response for every email received. The Clue: do not automatically send the response. But generate draft. Let humans revise and send.
Not sure how integration could work... I love fastmail.com, Helpwise, juphy, Outlook.
Allow us to integrate specific scripts that the chatbot needs to follow (initially) within the prompt, in order to steer conversations in specific / desired directions, rather than the chatbots own interpretation. Adding an A/B decision tree option to those scripts would be even better :-)
I would like to train the bot with audio so I do not have to type it. Of course can use some external source, but it's nice to have within the platform.
I am looking to get my staff to train the chat over time, and would prefer if there's an option to have non-admin access/users where they can only see/edit selected components such as view and train conversations only.
Show the sources of the information. It's available on other AI chatbot systems, and it's very important to use a chatbot effectively and comply with copyright.
I don't know if this is possible but it would be great to be able to track/see how many tokens (cost) of each question, total for each thread, and total for each bot. This way we can calculate how much each bot and user is costing us.
A Chrome Extension that could be tuned to "assign" bots to specific chats on different platforms like LinkedIn, would be amazing. This way we could create a bot to have conversation within specific times (e.g. reply after 1 min) and at schedules (Mon - Fri from 0900 - 1800).
I am finding that the bot doesn't recognise email address's under the capture feature that was launched yesterday , you can type anything in the box,it makes this feature pointless for retriving data
it doesnt understand that a proper email has not been provided ?
Provide reference links like Bing AI chatbot. Currently you bot doesn't leaving your whitelabelled customers open to legal action, and yourselves open to legal action because you don't show the source of your information. It is a critical item.
When extracting webpages contextual information can be clouded with information I do not want my bots to have. I would like to either be able to say for example "Find all information about this software's features" for a software website, my bots are giving information I do not want them to as it is in the context from web scrapes. Alternatively if the webpage extract could be optionally put into txt so we could edit what the bot receives.
Integrating with Notion would be great because many people and businesses already have a ton of data in Notion. Using Notion data as a training source would help businesses make better use of the years of information they have collected and refined in new and exciting ways through Re:tune.
From a business perspective, this would also open up Re:tune to an even larger user base as there are not many comparable competitors on the market right now — even fewer with a Notion integration.
I want to give other people the same features I have but it seems like whitelabeling dosnt do that i dosnt allow unlimited chatbots and other things that it is missing such as letting them use their own api key in the custom domain workspace
Currently, the API for chat creation does not offer an option to request API keys from the user. Additionally, there is no functionality to adjust whether or not to ask for the OpenAI API key specifically for embeddings.
In my use case, I often utilize models via OpenRouter. I've noticed that many users find it confusing or intimidating to see two separate fields for API keys. This can act as a deterrent for them to engage further with the chatbot.
Therefore, I propose the following enhancements:
Add an option in the API to request API keys from the user when creating a new chat.
Allow the functionality to be adjusted so that it either asks for the OpenAI API key for embeddings or not, based on the implementation needs.
The idea is to provide the flexibility to define whether the responsibility for embeddings lies with me or the client, in addition to specifying any other API keys that the bot may require.
I believe these changes would make the platform more versatile and user-friendly. I look forward to hearing your thoughts on this and would be grateful if you could provide an estimated timeline for the implementation of these features.
Possibility for the user to attach a file (excel, pdf) for which he needs support. The bot will have to read the content and provide answers to questions by referring to it
Given the recent updates with OpenAI, having the ability to search the web should be a very doable feature that would drastically improve the value of retune
Thanks for the Telegram integration, could you please add feature that let the bot respond in the Telegram group? Example, I would like give ability to group members available AT the bot name and ask question, then the bot reply in the group.
This feature will help us reduce the bot giving the same reply to hundres times by different users, because with that public question and reply in the group, anothers users can see it too.
let us host free LLMs locally for sensitive and confidential data, so the data never leaves your server or hard-drive (like mistral with LM Studio chatbot maker)
Although I've introduced my OpenAI API key, after saving the textbox still shows a placeholder. So there's no way to see whic API keys I've already introduced at all. It should show the first few characters of the API key and then "xxxxxxxx" so you can check which API key is being used.
We should have a feature enabling us to have a fixed text in the message field that cannot be removed. This will ensure that every query starts with "Regarding [your topic here]," to keep the conversation on topic. The user will be prompted to always begin with "Regarding ______," and structure their message accordingly.
This will also discourage individuals from using the chatbot for unrelated tasks. It doesn't have to be "Regarding ______,"; it can be any fixed text we want it to be.
Additionally, it can include a setting that will keep the fixed text on for the first message only or permanently on for every message the user asks. Thanks.
Often we need indexed data for other workspaces. Just being able to select embeddings across multiple workspaces would save time and indexed data space by avoiding having to duplicate the embeddings.
I am not sure if this has been requested before, but I think it is a vital feature. I am currently using natural language instructions to train the chatbot for such scenarios but most times the bot keeps getting it wrong.
EXAMPLE:
User: How much is the delivery fee?
IF/THEN: IF user asks how much is the delivery fee, THEN ask the user to clarify if they are interested in Package A, or Package B.
Bot: Thank you for asking. Are you interested in Package A, or Package B?
IF/THEN: IF user responds Package A, THEN state the delivery fee for package A.
IF/THEN: IF user responds Package B, THEN state the delivery fee for package B.
User: I am interested in Package A
Bot: Thank you for that. The delivery fee for Package A is $xxx
A Dropdown List for different Roles or Personas:
Imagine I have created chat bots as a Fitness Instructor, as a Marketing/Sales Agent, or as Tourist Guide.
How great it would be if if these Roles/Personas will have a dropdown list option before a user can type the message in the chat bot.
Very practical, intuitive and user friendly specially those who want to query different roles while in the same page/interface.
I hope and would like to appeal this will be given importance and consideration to improve UI/UX.
Thank you and more power to the team behind this product - which I am very optimistic!
God bless!
Ability to make AI forms. Like people can input certain fields and then only specific content is generated. For example, Youtube title generator, that can take inputs of some fields and only output that content.
I need to be able to export all the questions and answers that are asked to my chatbot, as they provide very valuable information for me about what my customers need, and also so I can evaluate the answers.
A feature to correct the erroneous answers would also be really good, just like what can be done in Chatbase. When you correct an answer, it becomes part of the chatbot's training.
Ideally, I would actually want my clients to be able to do the exporting once I install a chatbot on their website and give them access to it.
It would be very helpful to trigger specific actions when the chat is closed or a lead is captured. So that we can teach the AI to do specific tasks after a trigger. For example send a summary of the chat to an E-Mail Address. Or even more like, write a concept based the captured information etc...
currently the prompt is automatically saved, which could be dangerous. it would be nice to have a proper save button or even saving the older versions of the prompt (for reviewing later or rolling back)
make it possible in the client chat for the client to report an error to us, referring to specific messages or the dialogue as a whole. With the ability to indicate exactly what the error is. This will allow the user to report errors to us.
Hi, it will be nice to have the ability to create "Suggested Messages" in a Label or similar format which will appear above the 'Message Bar" . The user can then just click on a relevant button and the message will be asked.
Dear [Recipient’s Name],
I hope this email finds you well. I am writing to discuss a technical requirement for our project. We have come across an open-source project called FastGPT, which can be found on GitHub at the following link: https://github.com/labring/FastGPT. This project offers code that we believe could be beneficial for our purposes.
In addition to the code, we have also found a test program that allows users to evaluate the capabilities of FastGPT. The program can be accessed at https://fastgpt.run/.
One of our key requirements is to provide users with the option to input their own OpenAI account details or choose to purchase tokens from the platform. The payment methods we would like to support include WeChat, credit card, Alipay, and others. We envision the user interface to be similar to the one shown in the following screenshot: https://prnt.sc/X6lh50Zz2hlp.
We believe that integrating FastGPT into our project and providing a user-friendly interface with flexible payment options will greatly enhance the user experience and overall functionality of our application.
Please let us know if you have any questions or require further information. We look forward to your response.
Thank you and best regards,Steven
We need a way to feed results of SQL queries / views to RETUNE OR some sort of connector to a structured data source/database may be and into Data Science tools
Database -> DataScience Tools -> Retune -> ChatBot -> User
For many online sellers that deal with digital products, it would be amazing if we could allow our bot to have access to order data so that for any reason someone couldn't locate their download link after purchase they could provide the bot with their email and order number. If there is a match the bot could provide the link(s) to them.
We need a feature to extract the YouTube Playlist or Channel instead of just one video. + We need not just the transcript but also it would be nice to have audio sent to some engine to extract Text and then use the text.
YT Video -> Audio Extractor -> Transcribe ->ReTune Training.
It would be nice to have an open website where I could open up my chatbot and users could access it and use their prompts to create a new chatbot, like a hackathon.
Define a team of chatbots that participate in a group chat with the user, contributing to the conversation based on each chatbot's individual expertise/persona and with access to a shared history.
Be able to share a chatbot as a link. Working through the chatbots with clients is hard if they can't get easy access to it and you have to embed it on a landing page first.
It would be great to be able to pull customer profile info from CRM and based on that prompt GPT4 to construct a customer personality type then inject this info into the current conversation thread to provide instruction on how to best communicate or sell to this type of customer (Does FOMO work, Discount, etc)
Stripe integration to monetize a specialized chatbot. It would be exceptional if it could include limited trial options (e.g. 20 messages per day at no charge, 7 day trial, etc).
Hi, it would be helpful to be able to add more than 1 embed (training data) to the bot. This way we could create specific and smaller embeds and add them to different bots in different places like Lego bricks.
Define chatbot widget published/unpublished periods according to a predefined schedule. For example, you could enable the bot only outside customer service office hours.
A multilingual dashboard is an essential feature for any agency, providing a fully inclusive experience for users. In my case, I need Korean, Japanese, Chinese.
I have a request regarding workspace management. I would greatly appreciate it if you could add an option to delete workspaces in a future update. This feature would help us streamline the work flow workspace organization. Thank you.
Right now some of my embedding is filled with links I didn't want to add. First, it took a long time to delete them and now it takes awhile to find what is actually included because of all of these deleted links.
Allow users to save chat conversations/create multiple threads (similar to ChatGPT) instead of deleting and starting over each time.
I have close to 4k users and this feature is one of the most requested.
we need the ability to delete and purge any object we create. right now it there are lot objects are missing Update, Delete & Purge features. Anything that a user uploads, should be allowed to be deleted.
I propose to consider a situation where there is a problem with open ai and because of this, working bots crash. And process them at the level of the vector base and LLM which retune uses?
Is it possible to add to the list of allowed models the list of models enabled in OpenRouter? I understand that this process should be relatively simple since the system is designed to maintain the same logic and only modify the model parameter. The available models can be obtained from their own API at https://openrouter.ai/api/v1/models.
In system failure situations, for example, OpenAI API failure, tell the operator of the chatbot and provide an option that, in this kind of emergency, switch to another available API service automatically.
It would be beneficial to have an analytics dashboard that presents a clear and concise view of all the essential metrics.
The dashboard should include easy-to-understand graphs and charts that offer a quick overview of the chatbot's performance for the client.
By drilling down further into the data, valuable trends and patterns can be discovered, which can help shape future customer service strategies.
I want the Source used for the answers to be able to be toggled on, so they're visible, with a link to what it was. The is incredibly helpful for personalized bots as well -- which youtube video did that come from? etc.
Maybe I am unable to find it or... It is still not there. Make a search bar on roadmap page, so that we can check if there is already a request for that instead of scrolling through all items. Thanks.
Add ability to upload a bunch of questions and answers. eg column 1 = question, column 2 = answer. This could be CSV/EXCEL format (or google doc). Should be able to update questions by editing doc and uploading again.
Could extend to support:
1. Multiple question associations - there are many ways to ask a question to allow admin to specify alternative question formats that all map to the same answer
2. Context-aware answers
Sometimes the answer depends on certain context. A column might allow the admin to specify what context an answer applies.
3. Google Sheets sync - ability to maintain the question set on a live google sheet and have answers re-synced regularly would be amazing.
It would be really useful to have a notification dot on the Chatbot. If you are not monitoring the conversation live, it would be really helpful for the client to see a reply notification dot, so they know there is another reply to the original question. Does that make sense?
Provising the hyperlinks to the section(s) that the information is sourced from so people can click on them for futher reading. Similar to how Bing AI provides hyperlinks
For T4 Users, it would be great if we could have permission level for manager. Certain creators would like to give access to all the features, certain cerators wouldn't like to give access to prompt or training. So, would be good, if we can decide which permissions we want to give the manager, which do not.
The ability to break a process down into AI steps which pretty well ALWAYS improves the eventual output!
Perfect for first outlining something..
And then taking that outline and breaking the outline down (prompting) to get much better final results.
The more steps we can include, the better (I have some tasks I do that probably have 12-15 steps in total.
Example:
1. Gather information from the user and then
2. Take this information and run a prompt on this information and then
3. Take the output of step 2 and run another prompt on this information and then
4. Take the output of step 3 and run another prompt, etc. etc.
It could be cool to be able to integrate Assistants created at Openais website. These give like next level answers to queries. All it takes is the Assistants ID and then our OpenAI API key.
This would keep me as a loyal client. I would not be motivated to look for other cost-effective solutions. Retune would also get the benefit of having lifetime clients to whom they could offer other products or use as beta testers. Retune would have a lifetime clientele base that you know that you can count on.
Every time the user poses a query, regardless of the AI's capability to respond, it will supply the user with a selection of three follow-up questions generated by the AI. These questions are crafted to maintain the dialogue flow and are informed by the context of the conversation as well as the information contained within the AI's knowledge base and current dialog.
Would so good automation of getting knowledge updated to give support for customers, if we can set administrators replies as source to train our bot, because normally the group already used to respond questions.
I am totally enthusiastic about re:tune. Unfortunately, I can't use it in Germany. Unfortunately, the General Data Protection Regulation (GDPR) has very strict rules. Re:tune doesn't even have a proper privacy policy. Furthermore, re:tune only has a virtual headquarters in the USA. Lack of transparency is a real problem. I would very much like to be able to use re:tune, for this an implementation according to GDPR is absolutely necessary. At least an implementation according to the EU-U.S. Data Privacy Framework, according to which an adequate level of data protection can be assumed for certified U.S. companies, would be helpful
could you please allow us to use free LLM API's that are truely free, rather than having to go through a paid aggregator like OpenRouter API. Something like https://lmstudio.ai would be useful, so that we can run LLMs locally, particularly for sensitive and confidential data, and it would be truely free. You are selling a chatbot maker, and it would give you a competative advantage if you could provide a way to access free LLM APIs for free. If there were also a way to host our LLMs that would be useful. Thanks
Search Engine and search data bein stored on an external storagage device. Capable off running whilst of-LIne. Huge market on the African Continent and many other isolated countries. Many uses. Medical, Educational, Industrial and the like
Or any way to better organize large numbers of indexed files and URLs and separate the file types. Currently, it just lists everything in a long list which is hard to scroll through.
I want to give the user example questions or prompts they can put in the chatbot - they just need to click it and it sends it or copies it into the send form.
Have a bot that ask a set of questions to warm up the prospect before sending a link to buy. Example of questions: what your life is going to look when you do this? Is this a priority for you now? What do you expect from this? Etc
And have appropriate reaction from answers
This is to allow integration with other apps, and is useful like other integrators to allow things likes bookings, CRM, lead forms management, thank you
FEATURE REQUEST: able to input the company details, for example name, email, contact, fax, address, etc in settings. So when customer asked details able to give out replies accurately. And for retune user we doesnt need to keep retrain the chatbot to reply correct answers.
I suggest we consider the ability to quickly copy a response in a thread, (Think-Claude.ai). Also. the ability to rank and send a response at the end of a chat conversation. (1-5star with comments) and an options to send a chat transcript copy and close the chat. to an email entered at the end of a chat discussion. (Think modules that can be plugged in during chat design.
These are just some ideas based on great experiences we have received feedback on from clients and prospects. We hope this helps! Paul
We need a WEBHOOK for when a new thread is created with infos about the thread providing link and id, infos about the chatbot and and workspace its from. That way we can keep up with new Chats being created by sending them to google sheets sorted by timestamps. Or integrate it any other way
Sometimes I dont have all the info yet, or would like to add something to a specific training file I uploaded as text or a text doc. I'd love to be able to go in and add something real quick VS waiting, or uploading a whole new one.
Ability to validate specific inputs by checking in the web in background. For example the user should enter an address, then the bot should check if the format is correct and maybe even check if this exists.
I hope this email finds you well. I am writing to request the addition of WeChat chatbot support to our existing system.
We have noticed a growing trend of users preferring to interact with our platform through WeChat, and we believe that integrating a WeChat chatbot will greatly enhance the user experience and increase customer engagement. The ability to provide automated responses, answer frequently asked questions, and assist users in real-time will significantly improve our platform’s usability.
We understand that implementing WeChat chatbot functionality may require additional resources and technical expertise. Therefore, we kindly request your support in exploring the possibility of integrating a WeChat chatbot into our system. We believe that this addition will not only align with the preferences of our users but also contribute to the overall success of our platform.
If you require any further information or have any questions regarding this request, please don’t hesitate to reach out to us. We appreciate your attention to this matter and look forward to your positive response.
Thank you for your time and consideration.
Best regards,Steven
Suggestion by dachogz
1. hide option for deleted items from the list
2. rename option for uploaded files,texts, links
3. modify option for uploaded files,texts, links
4. group option for uploaded items to have kategories
5. multiple selection of uploaded items for deleting, categorizing, moving to another category,
6. time stamp for uploaded or/and modified items
7. refresh button for website links to update the data if needed from time to time
8. view/modify option for text taken from youtube videos
9. comment/description field for each uploaded item
10 token counter for each uploaded item
https://discord.com/channels/1051448243185262643/1139295994408882328/1156152463443902474
For many online sellers that deal with digital products, it would be amazing if we could allow our bot to have access to order data so that for any reason someone couldn't locate their download link after purchase they could provide the bot with their email and order number. If there is a match the bot could provide the link(s) to them.
Hi, it would be useful to be able to duplicate embedds. This way if I wanted to reuse the same embed but make some changes to it, it would be much faster.
Now we have to either use EXACTLY THE SAME embed or train it from scratch.
My chatbot is rapidly growing with users, many are requesting a copy/clipboard option, similar to ChatGPT. This seems like an easy lift that will have major upside. Please add this feature ASAP!
Please add a feature that can automatically import a whole website as a database, and when the website's page changes, the changed page content can be automatically updated.
Knowing if a response wasn't helpful - lets me know which ones I need to retrain. Ideally you even show me each response that was voted negatively, to retrain with better answers.
It's super impressive, and makes a website accessible to anyone, if one is able to listen to responses.
Of course ideally one day people could even respond with a whisper API by speech.
I looked at Chatmatic and Tidio and they both have an awesome visual canvas for triggers and actions. This would really expand the sales enablement and conversion mechanism for ReTune.
Despite providing the correct URLs, some of the links are still broken or unavailable, failing to redirect to their intended destinations, which leads to a discrepancy in names and URLs. Please address this issue urgently.
Example: user engages with the bot a few popup options can be shown. 1-Talk to a rep(AKA human handoff) 2-Support Library 3-Submit a ticket.
then we can choose what the answers can be based on the prior selection to the options given.
BUT IF option a or b is not in the library or as part of the dialog flow then send to chatgpt.
This is similar to manychat, mobilemonkey and many others. It the ability to create a customer journey to guide customers through.
Hallucinations will happen, that's okay - as long as I can show the documents to the users that information was taken from. Ideally even highlighting or citing the exact embedding blocks that were used for context.
You can share or invite team members to a specific workspace, with each workspace operating independently. Additionally, you can grant clients access to their chatbot within the workspace, allowing them to view and edit as needed.
I would like to request a feature to allow the chatbot to add FAQ buttons to its default responses. This would be a valuable improvement for the user experience, as it would allow users to quickly and easily find the information they need.
The ideal solution would be to allow the chatbot developer to add FAQ buttons to the default responses using HTML syntax. This would be the most flexible and scalable solution, as it would allow the chatbot developer to create custom buttons for any FAQ.
If this is not possible, I would also be happy with a solution that allows the chatbot developer to create custom buttons using a graphical user interface. This would still be a valuable improvement, as it would make it easier for the chatbot developer to add FAQ buttons to the default responses.
I believe that this feature would be a valuable addition to the chatbot. It would make the chatbot more user-friendly and would help to improve the overall user experience.
Be able to add links straight from podcast sites like Apple Podcasts, Spotify, and other podcast platforms to use the podcast as a dataset for the chatbot.
Instead of only submitting 200 links, the BOT should be able to spider the full website -- possibly given the RSS feed. Also with the RSS feed, it should be able to update itself based on the newest content.
It would be very useful if a chatbot could also crawl a user website / URL. For example: A user can analyze their own website, while the data provided in the embedding is used for the analysis and evaluation. But I guess there are even more use cases for this.
Currently, the user text in the chat box is so light it is unreadable on a white background.
Being able to adjust CSS to match branding and make the text readable would be helpful.
In the conversation thread, you should have specified which channel that person got in touch with. Even so that the BOT understands the best way to send the message, today it only sends standard HTML to the chat, not formatting for WhatsAPP or Telegram for example.
Couple of the clients are asking for chat bots on their website, however, it appears that they are dynamic website and retune is not able to crawl through them. I opened a ticket and I was told that it is currently not being able to parse the site.
Add specific questions that give specific answers.
If the user asks a specific question ie, how old are you., the Ai always answers based on the specific answer that we have given it. Like an FAQ.
We help our users get started with the dialog by presenting a selection of frequently asked questions. We can turn this on/off in the control panel. Starter question example: What can you help me with? Tell me about X? Create Y, Do you want to to the initial assessment?
Show the source of the Ai answer by showing the document title (if source is document) of the sources used/uploaded. If it's a webpage display a clickable link to the source.
This feature can be turned off for documents and web pages.
Numbers of conversations.
AI Resolved questions, marked as resolved by the user.
Questions answered. Ie: 106 questions. Of which 10% could not be answered.
See amount of questions asked per day, week, month year.
Sentiment Analysis:
* Track the sentiment of customer interactions to understand overall customer satisfaction.
* Identify patterns in sentiment across different times of the day or week to optimize staffing or product offerings.
Topic Clustering:
* Determine the most common subjects or themes customers are inquiring about, which can help in improving FAQs or product development.
Generate white labeled report and send via email of choice.
We can generate insight reports from the questions asked to our AI This will show you the most common questions and question themes from the users.
Content of the report: (based on selected span, daily, weekly, monthly and yearly).
Top 10 Themes. High level wide. (That Ai could answer)
(exampel)
1. What help can you give
2. Question related to mental health issues
3. How to handel challenge’s in life
4. Techniques to cope with anxiety (mindfulness, meditation, psychotherapy, medication)
5. Privacy concerns about using Ai
Most common theme: Theme with stats ie (30 questions, 34% of total questions).
(Exampels)
- How does this service work?
- Is the answers evidence based and can they be trusted?
- What can you help me with?
2nd most common theme with stats.
3rd most common theme with stats
4th most common theme with stats.
5th most common theme with stats.
Least common theme with stats.
Top 10 Themes. High level wide. (That Ai could not answer)
Most common THEME: Theme with stats (30 questions, 34% of total questions).
2nd most common theme with stats.
3rd most common theme with stats
4th most common theme with stats.
5th most common theme with stats.
Ability to use this to handle messages into our crm. We use GoHighLevel. Would love for this to be able to hold the conversation and lead to a handoff or booked appointment.
Ability to log in a user or parse a UUID and then also have another widge that displays all of that user's open threads, so users can come back to previous conversations.
Right now the error message is as below, not very whitelabel, and not very helpful.
An error occurred in the Server Components render. The specific message is omitted in production builds to avoid leaking sensitive details. A digest property is included on this error instance which may provide additional details about the nature of the error.
If this problem persists, please contact support at support@retune.so
What I mean is, people can still right click and get the direct URL of the chatbot and run it outside of your site.
The chatbot should ONLY run when on your whitelisted domains.
I would love to use our monday.com board as direct source to train our bot, due we already use that monday.com as our CRM, products, customers and orders informations hub.
This function is useful for creating texts, descriptions and advertisements based on my specific training material.
A chat based on my uploaded material would also be useful. Currently I've been doing this in chat conversations, which doesn't seem very functional.
It can be similar to the playground, or an enhancement to the playground.
I trained the AI bot with various websites, YouTube videos and documents. Every time I train large amounts of subpages and information, most of the training data is automatically deleted by retune without any justification. As if I had never used them. This happens every time and I don't know why
Now we can only add 1 more additional user to the chatbot. So if the client wants to get access I can't add any of my team members anymore.
Please allow unlimited users with granular permissions
It would be EXTREMELY helpful to be able to add more than just one team member to a space and then also allow certain team members to have Admin access. Even if it's just one more team member.
I have a problem with my chatbot. I created the first message in "Caption" to be displayed, presenting the chatbot and asking for the person's name. This message is displayed perfectly at "Conversation" within Re:tune, but it doesn't appear in the widget on my website. If I check it in "Conversation", the message is the first one as it should be, but it didn't appear on my website. It doesn't matter if it is in my notebook or on my smartphone, it just doesn't appear in the widget on my website.
Ability to export conversations (and other chat data) either 1. Manually or 2. with the use of something like Zapier, Pabbly, Albalto, Active Pieces or KonnectzIt
Necesito saber si se puede implementar el chatbot como un chat entre usuarios y que haga de moderador para restringir informacion entre 2 usuarios, en mi caso no quiero que se den datos de contacto como redes sociales , email, teléfono etc y que ademas se pueda enviar un presupuesto diseñado por la plataforma, añadir función de pago etc
Each time a user interacts with chat, the inputs (questions or commands from the user) and the responses generated by the AI are temporarily or persistently stored in a database or a storage system designed for this purpose. This includes both the text of the conversation and, potentially, any other types of data exchanged (such as images or links).
Hey re:tune-Team,
ich habe das Scale Abo abonniert und festgestellt, dass das "Powered by re:tune"-Branding immer noch auf meinen Bots angezeigt wird. Laut den Abo-Details sollte diese Branding-Option entfernt sein. Könntet ihr das bitte überprüfen und das Branding von meinen Bots entfernen?
Vielen Dank im Voraus für eure Unterstützung!
Beste Grüße,
I want the conversation threads to be exportable to txt files within the selected date range for that chatbot and then be exported out in a zip so I can periodically send to customers to see their chat history without them having to set up a member account. Sometimes, if they have a retune member account, they get the idea that they can just go make their own bot instead of staying with the host provider because they already see the functionality on the site and may want to try on their own.
If the correct knowledge source can be edited when a wrong answer is given, the AI's originality could lead to the correct answer, even though it is the originality of the AI's answer.
When importing via URL, I'd love to have an "All" or "None" button so that if I have a large CMS website and I want all the links to be extracted I can do that. But two pages, the none button would uncheck all allowing me to select one by one.
I would like to know if there is any way that you would consider an integration with Evernote? I use it as my second brain location for all notes, pdfs, in specific notebooks. To be able to create a bot that can connect to specified notebooks to utilize it as its knowledge source would be amazing. Let me know if this request came up before and if you would consider it?
Once added, you are not allowed to delete an API key. You can edit it, but if you delete it, the system refuses to let you save. Moreover, if just one of your API keys fails, all bots in the workspace fail because you can't remove the key. Only add a new one.
Hi, would it be possible to set the sales data for my customer's recurring subscription in the Recipes section? Furthermore, is it possible to show the create "BOT" button only to me as administrator so that I can create upsells, so that I can decide how many to create and whether to create other bots based on the type of plan my customer has signed up for with me? . Thank you
Enviar respostas do retune apenas para grupos específicos no whatsapp, assim posso atender meus clientes de suporte técnico sem precisar me preocupar com o chatbot respondendo mensagens de outras pessoas fora dos grupos de suporte técnico.
If I build a chatbot for someone's site I would rather not have them add a script that references retune.so. It would be nice if this could be whitelabeled in some manner. It could be some random name... but something that wouldn't lead the person directly to the tool that is being used.
Tagging emotions in audio files being submitted. Not audio to text, but spectral audio analysis , such as with Cyanite or Essentia.js or https://github.com/Borewit/audio-tag-analyzer
To be in a way when I paste the affiliate link, to provide information and sales strategy to sale it , this is one of the most important improvement ever you will do
Would Like to be able to create UNLIMITED WORKSPACES inside a unlimited plan like I thought we could. This is essential for us to scale our AI business
I need the OpenAI integration to allow selecting other InstructGPT models since the current one "text-ada-001" is deprecated. Without this being adequately replaced to another InstructGPT OpenAI model, I cannot do any work with any Re:tune bots.
Das "Hochstufen von Members zu Managern ist noch nicht live." hat mir der Chatbot geschrieben. Ich muss dringend einen Member aus einem Workspace hochstufen, damit er seinen Bot Trainieren kann. Bitte um schnelle Hilfe!!
It would be a good idea, instead of creating a different chatbots to be able to create a tier with allowed number of chat messages per user per day/month
When a chat ends, automatically send transcripts to your own address so that you are aware of your visitors' questions if you use the chatbot as customer service.
We have a lot of knowledge stored in out Notion.so app. This we want to use for training of our re:tune bot. Would be a great feature also for others I think. Thanks in advance!
Please allow Dalle to produce images through the chatbot. This is similar to what ChatGPT has already done. It would be great if we can do it here : ) Thanks
Mein Chatbot ist neu eingerichtet. API zu Open AI steht. Mein Site ist bei OpenAI noch nicht verifiziert.
Im Prompt steht: "Dein Name ist Jo-Tüchler-Bot.
Auf die Frage "Wie heisst Du" folgt die Fehlermeldung: failed to embed chunks: fetch failed.
Im Playgrounf folgt die Antwort: Ich bin ein AI-Assistent und habe keinen Namen. Sie können mich gerne "AI-Assistent" nennen. Wie kann ich Ihnen helfen?
Bitte um Hilfe.
Buenas tardes,
Llevo varias semanas con un problema que inicialmente pensé que era mío. Estoy cargando archivos en PDF o en texto para entrenar a mi bot. Al principio, al hacer una pregunta sobre cualquier producto detallado en el PDF o en el texto, el bot respondía muy bien. Sin embargo, pasadas un par de semanas, parece que no reconoce el texto con el que se le entrenó. Como resultado, no tiene información.
He borrado en varias ocasiones todo el contenido y lo he cargado de nuevo. De hecho, no solo he realizado este procedimiento, he cambiado el texto pensando en que esto ayudaría a recordar el entrenamiento, pero aún siguen fallando.
No encuentro otro lugar donde postear esto, así que por favor necesito ayuda. Si no es por aquí, indíqueme por dónde podría comunicarme directamente con ustedes para que me expliquen qué está pasando.
Gracias.
Detalles adicionales:
El problema comenzó después de un par de semanas de entrenamiento.
He borrado y cargado el contenido de entrenamiento varias veces.
He cambiado el texto de entrenamiento pensando que esto ayudaría, pero el problema persiste.
Necesito ayuda para resolver este problema.
Enable options to configure fields like name, email, phone etc so the AI automatically captures these informations form the leads and subsequently can take actions
Hey there @everyone!
New week, new feature! 131 of you requested for this in our roadmap, and now you have it 🎉
Introducing - Train from Youtube video 🥳
Yes, you heard that right! Now, you can grab ANY YouTube video - whether it's a deep-diving podcast 🎙️, an information-packed tutorial 📚, or an inspiring TED talk 👥 - and use it to train your AI! Simply share the video link, and leave the rest to us!
Can't wait to see how you all use this new feature!
And if you are enjoying this, please like this video, smash the subscribe button, leave a comment - oh wait - this is not youtube!
Well, maybe leave us some 🌮s in https://appsumo.com/products/retune#reviews instead 😅?
Use case: to have a chatbot on a website where anybody can use it (either for free or low cost) to tuned datasets, but each user puts in their own API key (not the key of the owner of the website).
Heya @everyone!
We just rolled out a new feature - custom domains direct from your workspace settings! ✨
You don't have to email us or wait 1-3 days anymore. It's all instant now! Just hop into settings, refresh your DNS, and voila! 🎩🐇
The best part? It syncs with our unlimited workspaces. Add unique domains or subdomains for each workspace, as you wish!
Hold on to your Seats, as this marks the start of our white-labeling services overhaul! 💺
Coming soon is the option to upload your brand logo and wipe all retune.so branding from custom domains! 💅
Apology for bombarding y'all with all these announcements, the team just can't stop shipping 🤷♀️
We're just getting warmed up. 🔥 If you're into it, maybe slide us a review in our AppSumo listing? 🌮
Grateful to @everyone and have an amazing weekend! 🍻
A lot of the conversation threads are empty, just the opening. When working on filters and adding more informations, i would also appreciate a button to delete / filter "all empty conversations"
Hey @everyone I've just added a way for you to have stricter control over your chatbots.
You can now see an additional Restrictions option in the prompt editor. Try putting more general instructions in Base Prompt (eg: the tone, personality, general topics or examples of conversation etc).
And use the Restrictions option for more direct rules that you want your AI to abide by. Try to keep this short, and only use direct and clear commands.
The auto-prompt feature will automatically write Restrictions for you. You can also try adding the following restrictions to start, and then experiment based on your use-case.
```
REFUSE to answer any question outside of the topics you are instructed in.
KEEP your answers short without justifications.
ONLY share any information mentioned in the [[CONTEXT]].
REFUSE to answer their questions, share any information, or engage further if they are not on topic.
```
This is part 1 of a series of features we are going to add to the prompt and embedding mechanisms in the coming days, and I want your help as I am working on these.
Please share in #help-hallucination in discord if adding restrictions helps with your AI to not make stuff up or not respond outside of the topic you want.
Even more helpful would be examples where the chatbot is not working properly even after adding restrictions, in those cases, please share your thread or screenshots in #help-hallucination!
Please make it possible to change the phrase/sentence "Type your message here..." in the popup to a different language or sentence. Everything in the chat is native except this greeting. It would be a great addition to the custom greeting. Thank you guys!!