Welcome Guest ( Log In | Register )

Outline · [ Standard ] · Linear+

 AI Chatbot development, How to choose open source platforms

views
     
TStechman007
post Nov 22 2024, 11:16 AM, updated 2 months ago

Getting Started
**
Junior Member
199 posts

Joined: Jul 2017
From: Kuala Lumpur
Hi there,

I'm looking to build an AI Based Conversational Chatbot mobile application, using voice (TTS, STT) and NLU capability and intelligence from pdf documents and LLM to answer citizen's questions.

The main question is, which platforms, tools, applications, are open source and what are your experience and tips on it ? Or I need to stick to commercial options, I know of Microsoft Azure OpenAI, or Copilot Studio etc. Or can use Google's for free ?

I want to keep the cost minimum, until I see the value potential of the application.

Thanks in advance!
redza2k4
post Nov 23 2024, 07:03 AM

New Member
*
Junior Member
29 posts

Joined: Oct 2009
Rasa or Botpress to build the chat bot if you’re on a budget, is supported for voice and is open source. For LLMs, Hugging Face is great, for extraction from PDFs Baystack is great!

Google and Microsoft’s are all commercial options and they’re fine for testing, but they can get expensive as you scale. The flexibility on no costs, but definitely more work on setup side with open source. Basically it comes down to having features and budget.
TStechman007
post Nov 23 2024, 10:46 AM

Getting Started
**
Junior Member
199 posts

Joined: Jul 2017
From: Kuala Lumpur
QUOTE(redza2k4 @ Nov 23 2024, 04:03 AM)
Rasa or Botpress to build the chat bot if you’re on a budget, is supported for voice and is open source. For LLMs, Hugging Face is great, for extraction from PDFs Baystack is great!

Google and Microsoft’s are all commercial options and they’re fine for testing, but they can get expensive as you scale. The flexibility on no costs, but definitely more work on setup side with open source. Basically it comes down to having features and budget.
*
Thanks for your valuable response and advice on this. Let me explore on this.. Any idea about LAMA, is also open source, how's its accuracy compared to OpenAI ? And Hugging Face would be similar ?
SUSzhiyung
post Dec 20 2024, 06:49 AM

Getting Started
**
Junior Member
133 posts

Joined: Jan 2003
From: Kajang



I’ve worked on something similar, and for open-source options, I’d recommend checking out Rasa for NLU—it’s super flexible and works well for conversational bots. For voice, Coqui TTS is great for text-to-speech, and Vosk works nicely for speech-to-text. If you’re dealing with PDFs, you can use LangChain to pull data and connect it to an open-source LLM from Hugging Face—some of their models are surprisingly good and don’t cost a thing if you have the hardware to host them. I’d avoid commercial options at first since the costs can add up fast, but Google’s APIs do have free tiers you can experiment with. Start small, keep it simple, and see how users respond before scaling up.


TStechman007
post Dec 21 2024, 09:54 AM

Getting Started
**
Junior Member
199 posts

Joined: Jul 2017
From: Kuala Lumpur
QUOTE(zhiyung @ Dec 20 2024, 03:49 AM)
I’ve worked on something similar, and for open-source options, I’d recommend checking out Rasa for NLU—it’s super flexible and works well for conversational bots. For voice, Coqui TTS is great for text-to-speech, and Vosk works nicely for speech-to-text. If you’re dealing with PDFs, you can use LangChain to pull data and connect it to an open-source LLM from Hugging Face—some of their models are surprisingly good and don’t cost a thing if you have the hardware to host them. I’d avoid commercial options at first since the costs can add up fast, but Google’s APIs do have free tiers you can experiment with. Start small, keep it simple, and see how users respond before scaling up.
*
That's a wonderful advise! Thanks for sharing your experience. And I assume you'd recommend Python for all of these products, would work best ? I'll look for sample code for each of these.
mrlive005
post Jan 22 2025, 09:24 PM

Getting Started
**
Junior Member
108 posts

Joined: Nov 2013
I've looked into AI chatbot development for my business, mainly for handling customer inquiries and automating basic support tasks. Open-source platforms like Rasa and BotPress are great options if you want flexibility and control over your chatbot without high initial costs. They offer solid NLU capabilities and can integrate with TTS and STT tools, but they do require some technical effort to set up and maintain.

If you're just starting, Google’s Dialogflow has a free tier that covers basic needs, and it’s pretty user-friendly. It’s a good way to test things out before investing in paid options like Azure or Copilot Studio.

For my business, I decided to go with a ready-made solution from https://logitize.ai/chatbot/ to save time and focus on other areas. It’s been helpful in streamlining customer interactions and improving response times.

This post has been edited by mrlive005: Jan 27 2025, 08:30 PM

 

Change to:
| Lo-Fi Version
0.0139sec    0.82    5 queries    GZIP Disabled
Time is now: 10th February 2025 - 11:55 PM