Demo idea: arrange with Google to do a dual demo where Rasa is talking to Duplex about an agreed upon topic.
Hello akelad, can we integrate rasa with whatsapp (through whatsapp business api) if yes , then please help me with it?
Bot example which can explain everything in just one bot. -
- Using buttons and sending custom payload using http webhook or other webhooks
- Using all webhooks at the same time on the different ports.
- Enabling and disabling CORS using cmdline
- Training models, adding intents using http.
- Bot having multiple custom actions (as in previous example there was only one custom action, I spent alot to figure out how to create second) having simple REST API calls to third party servers (GET POST DELETE)
- Some example having some multi user case.
These are problems I think every developer want to be solved at one place.
Thanks!
2 Likes
A bot that accepts data from user in excel format. A live demo of dynamic data training would be really helpful
5 Likes
Bonjour @chrissofvista ! Je travail sur un projet qui consiste à mettre en place un chatbot pour les services public qui permettra de répondre aux questions concernant toute procédure administrative comme demander un extrait de naissance ou comment procéder pour obtenir un bulletin de naissance… J’ai commencé avec dialogFlow mais j’aimerais utiliser Rasa. J’ai déjà commencer mais mon modèle Core ne gère pas bien les contextes. S’il vous plait j’ai besoin d’aide. Si vous êtes d’accord, je peux vous envoyer mon mail et on pourra échanger. Toute aide est la bienvenue @andreasstuber, @Juste, @akelad, @mahdimor …, car je suis novice sur cette technologie qui me passionne beaucoup. Excusez moi, je suis francophone et j’espère bien que vous comprenez le français !
Bonjour, pour une meilleure gestion du contexte il faut en générale ajouter des stories. Sinon changer le modèle de prédiction. Je peux te donner un coup de main si besoin.
Sylvain
2 Likes
Super @RobbianoSylvain! Comment puisse vous contacter pour qu’on en discute ?
Bonjour @nabzina, le projet m’a l’air très intéressant! Quoique pour avoir des réponses rapides, je pense qu’il serait préférable de poser les questions dans les sections appropriées. P. ex. essaye de poser ta question concernant Core ici. Si tu la formule en anglais tu auras certainement plus de réponses ! 
1 Like
Ok @Saladdin 
merci
Rasa Core must introduce a NLU bypass stub . Which will enable the bot to bypass the nlu whenever it sees that stub.
For example :
- intent1
- do something
- intent2
- do something
- bypass_stub
- do_something
when the bot sees bypass_stub it ignore nlu and takes the input irrespective of intent or entity . This feature will be a gem as there won’t be any necessity to create data for rudimentary inputs and also for long feedback or any kind of data for which building an NLU model makes no sense
RASA TEAM THINK ABOUT IT
3 Likes
Kindly mind the choices of words that you use, MUST is not a good one, but your idea makes sense though.
3 Likes
@Abir something like this should already be possible. If you send a message starting with a / it will be directly interpreted as the intent without an NLU classification, e.g. /greet will result in the bot logging a message with intent greet and Rasa will skip the NLU classification.
1 Like
HELLO
Is there a better way to help the RASA community than teaching the AI agent to advise users on problems they can face? Just use the community QA to feed the bot. Ppl will find it useful and it will free up some of your resources.
Peter
A bot to make Rasa self-hosted. In other words, a bot to make new Rasa bots, that interactively acquires data/configuration/customization from the user and gives him/her a bot in the end 
.
1 Like
Is there any method for bot to understand the phonetics or metaphor words and get the exact entity which i would have listed. The Synonyms method is ok, but what if the user type any other language in English words.
I hv cmpltd my enginering in Anna Univrsty.
The actual reply of user would be
I have completed my engineering in Anna University
The highlighted words are entities. Can we able to do this in existing flow, if not is this possible to do in rasa.
Thanks…
3 Likes
I found a new thing “How to revoke the conversation?” as the example, if multiple users did conversation with model and one of user busy for 5-6 hours and after that, he again starts the conversation where he left the conversation So how we revoke the conversation that time and how to take input for revoke conversation? You have Any Idea… How to solve this issue… I also need this… and How to manage this… ?
maybe you can use fuzzy matching for this. but i’m not sure how accurate it can be.
I have a new Idea ,That is the Auto populated questions. It means user asked questions to bot then related questions also suggested like drop down. For example user asked “hi” then some suggested question like "hey how are you ,how can i help " etc…This will help for user what kind of topics we have.
2 Likes
hello If i give any query it should not do nlu classification some times .Why means,suppose if bot ask please enter your name then, user enter name then .Then nlu classification for given name will get less threshold value .So i request ,If their is any solution for this one . can u please reply back.