It would be great if the ai agent can build the tool when requested rather than coding it manually and if there is a complexe tool linked to a service then it should ask for its credentials. So basically making an ai agent that is specialized in providing the tool when requested
oups working without api key is needed you do not give many ollama models yet or do we have to make an agent to exec bash => conda => ptaisouai env => another task with else model (not very efficient) same for many producers (including free) for example nexa sdk or genesis continue your marvelous work Mervin
Good morning, My friend I would like to know the following, I have the information that the agent under the hood he makes many requests, like OpenAI and deposited this account it arrives very expensive because as it does a lot of things that we do not know to abstract the big code there are things that it does and we do not know that proceeds
I would like to know this information that for me is crucial to implement this fremework in my application because I don't want to have surprise in the OpenAI account or other llms, by fsvir answer me and help me
Please check your code. I cannot run it properly with lm studio. Ui insists on requesting api key although I have already set correctly the .env. What runs correctly is only runing the “praisonai” with the agents already set for the mars script. Please stop replying only to kind words and give sone help, otherwise the video and code are useless. Sorry for bit€h!ng but I have lost 4 hours trying to make it run.
Okay. You tried to pulled out of something, which is alredy exist in many forms.. TogetherAI-s MoA-architecture is opensource and the fastest.. Stop fooling your audience. I Unsibscribed. Havea good one!!
Some feature might look similar. But here the ultimate goal is to make it simple and easy for any one to create and run agents. Also with transparency . All these are OpenAI sdk calls so it can be easily traceable. I’ll explain how to trace all these calls in another video.
@MervinPraison I'm sorry. I spoke harshly and judged instead of appreciating your efforts. It's not like me to be this critical. I'll take a closer look and provide constructive feedback this time.
It would be great if the ai agent can build the tool when requested rather than coding it manually and if there is a complexe tool linked to a service then it should ask for its credentials. So basically making an ai agent that is specialized in providing the tool when requested
thanks , for the video! what is the difference with Pandiatyc AI ?.. Thanks and greetings?
ABSOLUTE GOLD!!! WOW
Thank you
very thanks, you better than swarm of openai ;)
Hey Praison, can we use Open Source models like Groq or Gemini?
Yes you can!
Works with functions calling?
Yes
So PraisonAI UI is already set to run off of Praison 2.0, correct? Nobody wants to work out of Terminal :(
Thinking of integrating to UI effectively
top top top top 👋👋👋👋👋👋👋👋👋👋👋
so these are your agents you made? Looks very impressive
oups working without api key is needed
you do not give many ollama models yet or do we have to make an agent to exec bash => conda => ptaisouai env => another task with else model (not very efficient)
same for many producers (including free)
for example nexa sdk or genesis
continue your marvelous work Mervin
thanks :)
How to make it Claude-compatible?
Awesome
This looks great. Can you do a link to the specific code I could not find in your library?
github.com/MervinPraison/PraisonAI/tree/main/cookbooks/tools
Does it work with different Open AI compatible providers?
Yes
Works locally with ollama without openai api key?
Yes
Good morning, My friend I would like to know the following, I have the information that the agent under the hood he makes many requests, like OpenAI and deposited this account it arrives very expensive because as it does a lot of things that we do not know to abstract the big code there are things that it does and we do not know that proceeds
All it needs is an better ui throw it react and it should do good
Yes considering that and working on improving
I would like to know this information that for me is crucial to implement this fremework in my application because I don't want to have surprise in the OpenAI account or other llms, by fsvir answer me and help me
👍👍👍
Please check your code. I cannot run it properly with lm studio. Ui insists on requesting api key although I have already set correctly the .env. What runs correctly is only runing the “praisonai” with the agents already set for the mars script. Please stop replying only to kind words and give sone help, otherwise the video and code are useless. Sorry for bit€h!ng but I have lost 4 hours trying to make it run.
on Mars not in lol
Okay. You tried to pulled out of something, which is alredy exist in many forms..
TogetherAI-s MoA-architecture is opensource and the fastest..
Stop fooling your audience.
I Unsibscribed.
Havea good one!!
Some feature might look similar.
But here the ultimate goal is to make it simple and easy for any one to create and run agents. Also with transparency .
All these are OpenAI sdk calls so it can be easily traceable. I’ll explain how to trace all these calls in another video.
@MervinPraison I'm sorry. I spoke harshly and judged instead of appreciating your efforts.
It's not like me to be this critical.
I'll take a closer look and provide constructive feedback this time.