Name: PraisonAI
Uploaded: 2024-12-19T11:34:30.128Z
Duration: 2 min 33 s
Description: An open-source, low-code framework for building and managing multi-agent LLM systems, emphasizing simplicity, customization, and efficient human-agent collaboration.

praisonai

An open-source, low-code framework for building and managing multi-agent LLM systems, emphasizing simplicity, customization, and efficient human-agent collaboration.

PraisonAI

This is amazing introducing PraisonAI. You might have a question which tool to use whether to use crew AI or autogen. You might need to learn many functions to create multiple agents. You might need to create many files to orchestrate those agents. Confusion on which tool to use. Many functions to learn. Many files to create. 

What if we have a solution where you have one tool to choose, one file to create, and one function to learn? Introducing prais and AI. This application combines autogen and crew AI or similar framework to come. It is a low code solution. Just one page of text is more than enough to create multiple agents to perform a task just like this. 

And also it is in text format. You don't need to write any bit of code. If you want to use crew AI just say framework crew AI. If you want to use autogen just say framework autogen. As simple as that you are able to access user interface using prison AI hyph hyphen UI. You are able to deploy directly to Google Cloud platform by just adding hyphen iph deploy to the end. 

Your Lodge language model can automatically create multiple agents or create a team based on the task you provide automatically. At the end you will get the output. That's exactly what we're going to see today. Let's get started.

[Music] Started. Hi everyone. I'm really excited to show you about PraisonAI, a low code solution which I created to run multiple agents. Autogen and crew aai being the back end. This is like an o trator. In this, we are going to see how to install this tool, how to create AI agents, how to create user interface, and finally how to deploy directly to Google Cloud. 

I'm going to take you through step by step. But before that, I regularly create videos in request Artificial Intelligence on my YouTube channel. So do share AI Agent Store with colleagues. Make sure you click the like button so this video can be helpful for many others like you. 

So first step, pip install praon Ai and then click enter. Once after it gets installed, export your open a API key like this and then click enter. In this, we are going to see how to integrate this with open AI Gro and O Lama. 

So first we are integrating open AI. First to initialize, type PraisonAI hyphen iph init. Then you can type whatever task you want to perform. Create a movie script about cat in Mars and then click enter file agents.

Yaml created successfully. That's it just this one line of code and it automatically generated a agents file. I'm going to open this inside that file. This tool automatically created all the necessary agents: the idea generator, the character designer, and the screenwriter. You can keep this or you can even modify based on your requirement. The minimum requirement is to have one role, a backstory, goal, and then one task. That's it. Even if I delete all the other roles and keep only these lines, still this is going to work. But I'm going to keep the previous code with all the characters. 

Now the one function is also ready: one file, one function, and one tool. So here you are mentioning framework is Crew AI. Now I'm going to run this code in your terminal, Phasen AI, and that's it. Just click enter. The tool does all the heavy lifting and you can see the first agent started working: the idea generator. How simple is this? 

Next comes the character designer. The first idea generator created the plot, the ACT key. This script is about Gizmo, a domestic tabby cat. So the character designer is designing the character Gizmo: physical appearance, personality, background. Similarly, other characters in this script now it goes to the screenwriter and you can see a proper.

Script is being written here. External daytime Gizmo and Gizmo is in Mars and the whole story goes and finally it's completed. 

Now next let's modify this to run autogen. I'm going to change this to autogen. That's it. Now I'm going to run this code starting a user interface. So just type PraisonAI hyphen hyphen UI and then click enter. This will automatically start the grade your user interface and here is the user interface. 

Now we can ask create a movie script about London and New York. You can choose which framework to use. I'm going to choose autogen now. Then click submit. This will automatically create all the required agents and then start running. 

Finally, we'll get a response in the back end. We can see autogen already started the work. In our actual folder, we can see a new file got created called test.

Yaml just open it. There you can see the list of Agents: research specialist, storyline Creator, screenwriter, and it's about London and New York. Finally, we got the response here, so you can see the movie script about London and New York. 

So the AI agents got automatically created. They worked together, and finally, we got a response here, as simple as that. Now let's try also with crew AI, choosing crew AI from the framework and then click submit. 

And now we got the answer from crew AI. This can be confirmed when you see the backend logs. So the research analyst is there. You got the plot developer, screenwriter, and finally, we got the output. 

Now let's see a full automatic mode from terminal. To do that, type Praison AI hyphen hyphen Auto create a movie script about Tom and Jerry. That's it. You literally provided hyphen hyphen uto and what you want to do, and then click enter. 

Now it's creating or designing the structure, the number of Agents required, and their tasks, and the project got started. You can even see in the test.

Yl file the list of Agents it is going to create and what task they going to perform and finally we got the script for Tom and Jerry as simple as that. Now you can try with this and provide any task and try it. 

Next let's see how we can deploy this. Now I'm going to show you how to deploy this in Google Cloud but soon we'll be releasing AWS and Azure Cloud. 

So first step make sure that you have Google cloud account. Then you need to initialize using g-cloud in it. Next we are going to export open AI model name whatever model you want to use open AI API key. This is where you provide the API key. 

Then export open API base this is the base URL. This step is optional because we have already exported our API key before. I'm showing this so that you can customize this for olama for Gro and for any other large language model service provider. 

Once after you do this click enter. Now you have already created a file called agents. yaml and this is the agents.

Yaml playbook file and these are the agents we are going to deploy in Google Cloud as API. So whenever you call that endpoint automatically you get a response from these agents. You can also add your open model name API key API base in the EnV file. This is optional. So either you export all these variables in the environment or add that in this EnV file. 

Now just one command pron AI hyph iPhone deploy and that's it. Click enter. This will automatically build your Docker image, push that to Google artifacts, and run the image in Google Cloud run. Everything is done with just one command and finally you get the endpoint like this. 

Now I'm going to open this URL. Here is the endpoint and we got the response from the agents. Now you have successfully deployed the agents application in Google Cloud. This is just the beginning, but you might need to secure this endpoint as soon as it gets deployed or immediately go to Google console Cloud run, select the service which got published, and click delete once after you use it. If you want to permanently publish, then secure that endpoint. 

Now let's see how we can integrate ol LM Studio text generation webui and any other. So export your open AI API base like this and then click enter. Then export your open AI.

Model name like this I'm using mistal and then click enter. Make sure you've downloaded mistol by using olama p mistol. If anyone don't know about olama it is the easiest way to run a large language model in your local computer. You can download that from ol.

Now it's all ready. Now I'm going to type pr. a and then click enter. This will automatically use the agents. Yamal fall and start performing the task. Similarly, you can connect to fast chat with this endpoint LM studio with this endpoint mral API with this endpoint. 

Finally, I want to try with Gro. So exporting the model name, exporting open a API base with this grock URL. Next, export your grock API key in here and then click enter. Now let's run this pr. a and then click enter and it's going to be super fast. Within few seconds, it's all done. 

To integrate pra AI with any other python application, it's very easy. Just import PraisonAI and then you can call the function like this. Just provide the agents.

Yl file that's where all your agents are. Similarly, you can mention the framework to choose which framework you want to use. If you want auto mode, just use Auto and then enter what you want to create a simple function to integrate this with any other python application. 

Finally, these are going to come soon. Going to add tools hierarchical deploying to AWS and Azure. But to do all of this and much more quickly, I need your support to contribute and develop. I've provided the list of steps here. It's easy to follow and do. 

Let me know in the comments below what do you think about this tool and what you are expecting from this. I'm really excited about this. I'm going to create more videos similar to this, so stay tuned. I hope you like this video. Do like, share the AI Agent Store with colleagues, and thanks for watching.

Note, that there may be slight differences in text and video. Watch video for the original content.

PraisonAI Video