Praison AI UI: Automatically Create AI Agents for Work Automation
ฝัง
- เผยแพร่เมื่อ 11 ก.ค. 2024
- 🚀 Welcome to our Praison AI Tutorial! 🚀
In this video, we'll dive deep into the Praison AI user interface and explore how to integrate it seamlessly with Groq and LMStudio. Praison AI UI: Automatically Create AI Agents for Work Automation. This tutorial will help you boost productivity by automating repetitive tasks and creating custom AI agents tailored to your needs.
Tools: docs.praison.ai/tools/
What You'll Learn:
Integrating Praison AI with Groq and LMStudio: Step-by-step setup to streamline your workflow.
Creating and Modifying AI Agents: Learn to add custom tools and enhance agent capabilities.
Using the User Interface: Navigate through the UI to store conversations, change endpoints, and more.
Real-World Applications: See practical examples like designing a fantasy world and researching lung diseases.
Benefits:
Increased Productivity: Automate repetitive tasks and focus on what matters.
Customizable Agents: Tailor AI agents to fit specific tasks with custom tools.
Seamless Integration: Easily integrate with other AI platforms for enhanced performance.
Steps Covered:
Installation: Quick setup with pip install praisonai[UI].
Configuration: Secure your setup with a secret key and API integration.
Demo: Live demonstration of creating and running AI agents.
Customization: Adding tools like the internet search tool for advanced tasks.
History Management: How to log and resume previous tasks for continuity.
🔗 Relevant Links:
Patreon: / mervinpraison
Ko-fi: ko-fi.com/mervinpraison
Discord: / discord
Twitter / X : / mervinpraison
Sponsor a Video or Do a Demo of Your Product: mer.vin/contact/
Code: docs.praison.ai/tools/
🔔 Don't forget to subscribe and hit the bell icon for more AI tutorials! 🔔
👍 Like, share, and comment if you found this video helpful!
Timestamps
0:00 - Introduction
0:23 - Creating and Modifying AI Agents
2:00 - Navigating the User Interface
4:00 - Practical Examples: Designing a Fantasy World
5:30 - Using the Internet Search Tool
7:00 - Switching to Auto Mode
9:00 - Managing Chat History - แนวปฏิบัติและการใช้ชีวิต
This is great, a quantum jump from the available tools in the market.
Well done Mervin, this is impressive.
I run into a problem but got really fast response via Discord. Much appreciated Mervin ! Keep up the good work.
This is awesome. Thanks Mervin once again. Keep up the good work and sharing.
Damn Praison. This is great! Thanks for sharing all these.
This is amazing! I tried already to run this software, finally i have an guide.
Awesome Mervin! Exactly what i was thinking/looking for!
PS I like the streamlit version even better, but loading the various LLMs and keys etc seems a bit buggy. Is this just WIP?
Well done Mervin it is pretty awesome
Dude, great work!
I love your work abstracting agents.
Good work Mervin!!!
Nice! That really is a good starting point that leads to more productivity.
1 word: AWESOME !!!
Whoa nice work Mervin
This looks cool! Let me try. Wonderful channel, content and to the point, cut-the crap stuff!
congratulations Amazing work , really helpful for developer community and novice developers as well, indeed you are doing great work
Wooohooo!!!
1) Can we save these as some kind of prepared workflows for later? So if a similar request is being done, that it is reused?
2) can you please implement the code snippet to call it from terminal/python?
3) any plans to save history locally?
Rather amazing!!
incrível, top top top 👏👏👏👏👏👏👏👏
Mervin, good job!
Awesome 🙌🏻
very nice love it
thanks for the video, so nice...can use with any other model without openai key?
Perfect tutorial!
I wonder how to leverage this code and productionalize this. As in generating the entire crewAI code or Autogen Code
Mervin, you should mix it with this village UI we would see the face of the agents talking to each other in pixel art
Great Idea! :)
@@MervinPraison it could even grab a random avataaars and then create a "zoom" UI where you see them with text bubbles. I've been dreaming to do it since a while. I'll try to share you the mockup on Discord
@@sophiedelavelle5958 Sure. Thank you :)
@@MervinPraison I added you in MP (boloshon)
@@MervinPraison how can one add praison backend to work with flask or django project
🏆 Cheers for sharing!
It would be helpful for your non-dev viewers if you created a video on how to follow alone and try these examples with platform providers online. Whether we don't have the hardware or aren't allowed to use the hardware at work.
@Mervin, awesome work! @Anyone, did you experience problem that the output from the task is not shown even though the task is completed?
yes I have the same issue. I can see the results on my terminal but not on the UI
I did get the output back for a simple task so maybe it is hitting some sort of limit with my computer specs or something
@@rogercampbell9019 it's not working for me at all. Did you use crewai or autogen?
i dont understand the outputs. When using multiple agents, i expect to get one output per agent. The global output should be a menu to access every agent output.
Or is it one single ouput generated by multiple agents, like layers applied before getting one final result ?
Do I have to init something first? I just get an error: UnboundLocalError: cannot access local variable yaml_content.
Hey merwin, can we use gemini api instead of open ai api?
can we create better rags with it ?
great job btw
Is there a more UI based tool that you don't need to type tools as programming language?
cooler scheiß best i have seen echt gut gemacht weiterso 😁, dankeschön
plz make an addbutton for tools and the toolsadminagent can order by agents
Hello sir, hope you are doing great. I was not able to run the app in auto mode though it is generating agents.yaml file but not giving the output on the UI. While running in manual mode with same created agents.yaml file, it is giving output to the terminal but again UI part is only showing as loading.
I ran everything as you did from a Raspberry Pi5 (I had to create a virtual environment).
When I logged in for the first time (admin admin), I got Auto with the Chainlit logo with the words "str expected, not NoneType".
What is that?
Is this because I have not yet provided my API key?
On a separate note, on the left panel it says "Conversations are not persisted."
Small bug in there, i will fix it soon
For now first,
export OPENAI_API_KEY=xxxxxxx
then second, run
praisonai ui
That should fix this issue.
🥳🤯🥳💥
Wil this work for windows users?🤔
Tools? Awesome brother, but i cant find anything in the description, on the tools
docs.praison.ai/tools/
Your docs for tools does not have working hyperlinks so I can’t see the code for the tools
Can it perform tasks on my pc? Like working on my websites and running my ecom businesses by responding to emails and deleting junk mail etc
A real virtual assistant
@MervinPraison how can one add praison backend to work with flask or django project
nice, can you add your own knowledge base and do RAG?
all these will work on only models with function calls right sir ?? shall we use gpt3.5 trubo?? pls tell me
This tool requires function calls.
You can try gpt3.5 turbo, for basic tasks it should work
I get an error when running
$ praisonai ui
" 'tool_name' is not defined "
What can I do?
Can you please try upgrading and let me know if the error still exists.
pip install -U “praisonai[ui]”
So great! Thank you! 😀 But I cannot find the tools... even I cannot edit the agents and task definition. It whould be great to integrate Ollama as provider. Thank you nyway great project
docs.praison.ai/tools/
Ollama Integration: docs.praison.ai/other_models/
@@MervinPraison The tools link opens, but cannot see the code of the list tools.
how to have persist Conversations
I get a message that I've installed PraisonAI version 0.05, and praisonai 0.0.5 does not provide the extra 'ui'
strange since it looks like you are only up to v0.0.37? @mervinpraison
Did you try upgrading
pip install -U praisonai
Then can you please try
python -m praisonai
I need a WhatsApp agent.. tired of whatsapp chatting, but need to keep in touch.. need to post random stuff .. better would be to read and respond(which I dunno is possible🧐)..
is that possible with yr tool? Read agent, then a write agent..
where is the tools?
docs.praison.ai/tools/
install instructions are a bit "spartanic" and the experiment failed.
Pretty much this
Can you please try this
pip install "praisonai[ui]"
export OPENAI_API_KEY="Enter your API key"
chainlit create-secret
export CHAINLIT_AUTH_SECRET=xxxxxxxx
praisonai ui
Also if you have already installed this package before make sure you Upgrade by using this command
pip install -U "praisonai[ui]"
Also another option to run this is
python -m praisonai ui
@@MervinPraison it keeps giving error triead all--/usr/bin/python3: No module named praisonai
Probably it’s using a different python environment for you
Try
python -m praisonai ui
I guess every job all that it does is google search and produce reports ? 😆
im on a macbook . the input - praisonai ui- does not work for me i get this error. File not found: /Users/darth/ui
Does it need brackets? []
Dude use CursorAI and chat with it to have GPT4 help you debug
You might need to upgrade praisonai
pip install -U "praisonai[ui]"
Above command will upgrade it