I clicked so quickly on this! Really love your framework and how you keep dedicating time to improve it. I hear of autogen, crewai etc, but it looks like people are not seeing how good your framework is!
Your framework is really amazing! If you would, future video(s) explaining how to implement your agent swarm idea "for dummies" would be really helpful to a large amount of people who dont understand more "technical" concepts.
00:06 Genesys Swarm enables easy creation of AI agents without code 01:42 Agent Swarm creates Market Analyzer with new export file tool 03:12 The OpenAPI creator creates schemas from web pages for agent swarms. 04:36 Agency creation with API keys 06:02 Using various agents to analyze current Bitcoin market trends. 07:30 New updates and features in Agency Swarm 09:01 Steps to create and run Agent Swarms 10:25 The open-source AGI framework empowers community collaboration.
I would love to get a general idea of how much these swarms cost to run. Could you post a few general ranges of cost for your current tests? Nothing specific, just a ballpark idea.
Wish i could follow your install instructions. Im that dumb with 'code' etc. But have MANY IDEAS for 'AI Swarms'. If you know of any video i should watch please suggest a basic one which would help. Excellent video and excellent 'swarms'. Well Done. Frustrated me (cant follow install instructions) but EXCELLENT WORK Dude. Thanks.
Think about a way the major agents instruction code could be autofinetuned based on real performance. Like read your prompt, read your output, how your prompt could be updated to work better, self update your prompt. Or the senior Agent does this.
Yes great idea. The challenge is in allowing agents to measure the performance of other agents. Without this it’s hard to know how to modify the prompts
It could be implemented like asking the agent that is doing the task to get his own metrics and suggest improvements for himself and then the boss to implement it.@@vrsen
Great idea, but I think the problem with limited capacities and especially a limited number of repeating user-questions is that you need to repeat those measurements a lot for the same and slightly modified questions which you can't do in a working environment because a lot of offered prompt modifications will lead to worse output, which you don't want under no circumstances in your productive setup (because you already put a lot of work into the system prompts like they are ;-).
Imho modifying finally good working system prompts are not worth of trying to optimize them more and more to gain speed. But investigation the multiple steps (agent calls) a framework takes to solve more complex tasks and then trying to reduce the steps, that will increase the overall speed an save you a lot of money if you are using paid APIs. And what also may speed up complex tasks is giving the agents tools that can do more than one simple thing but maybe two or more related tasks which may be called by parameters, with arrays as output. So instead of calling multiple agents after another to each of them get one value which the next agent processes, maybe the "tool" may get all single required answers at once, give it back and only one last agent call may aggregate all that data and present the final answer. The tools could also have a parameter with which you may say that the tool will use the gained info in a pipeline and return one value or get the answers parallel an return an array. - Using async calls in the tool when it has to generate multiple parallel answers will speed up again. But it probably really needs a very well thought out and tested system prompt for letting a model decide if and when to use a tool in parallel or pipeline mode. Ouuuu... it's all so challenging :-)
Nice work. Question. How does agency-swarm compare with and interact with other similar projects like AutoGen, ChatDev and CrewAI? Are they able to work together?
do you know if it's possible, instead of using openai api key, to use local models with litellm's api wrapper (i think that's what it has)? i will try it myself soon but just wanted to know if you had any idea
I see your platform is based on OpenAI. I hear lots of talk about using LOCAL LLMs (even customize them), but so far I couldn't find evidence that it actually works. It could save money, allow for more privacy guarantees on local data, and make you less dependent on OpenAI's evolution (at least for some parts). Do you see any use?
# Chat with an intelligent assistant in your terminal from openai import OpenAI # Point to the local server client = OpenAI(base_url="localhost:1234/v1", api_key="not-needed") history = [ {"role": "system", "content": "You are an intelligent assistant. You always provide well-reasoned answers that are both correct and helpful."}, {"role": "user", "content": "Hello, introduce yourself to someone opening this program for the first time. Be concise."}, ] while True: completion = client.chat.completions.create( model="local-model", # this field is currently unused messages=history, temperature=0.7, stream=True, ) new_message = {"role": "assistant", "content": ""}
for chunk in completion: if chunk.choices[0].delta.content: print(chunk.choices[0].delta.content, end="", flush=True) new_message["content"] += chunk.choices[0].delta.content history.append(new_message)
# Uncomment to see chat history # import json # gray_color = "\033[90m" # reset_color = "\033[0m" # print(f"{gray_color} {'-'*20} History dump {'-'*20} ") # print(json.dumps(history, indent=2)) # print(f" {'-'*55} {reset_color}") print() history.append({"role": "user", "content": input("> ")})
I haven't been able to get the web_browser_agent to work with the genesis_agency. Am I supposed to start that first or is that a part of the genesis_agency jupyter notebook?
Hello mate! This is an amazing project! Congratulation on doing this; I've runned it on my side and I was trying to connect my agents to Mongodb, salesforce and AWS S3 bucket but the agents can't connect to it even with the appropriate keys given to it; Do you have any reference about how can I make it works? Any help is really appreciate.
Wonderfull, but I guess if there's any method to reduce the token request speed, as i am continuously running into "OpenAI Run Failed. Error: ', 'Rate limit reached for gpt-4o." I am also guessing if there's any way to select other models, such as gpt-3.5-turbo (apart from modifying the Python code underlying Genesis library, which I've tried to do, and btw didn't work)
Hi @vrsen, how could I increase the skills of each agent? I.e. I have one article writer + one article editor. Let's say I want to train the article editor by providing 3,000 words of examples of edited text. What's the best way to do this? Many thanks
This is extraordinary work. Congratulations. I agree with the final remarks around AGI. I really hope that Azure can release an Assistants API equivalent!
Hello! Nice work, and I'm really exited to build my first Agency Swarm! Do you have a community where we users can help each other? I've succeeded in installing Agency Swarm, but I'm having trouble understanding how to create tools using an Open API schema. I have the schema json file. One of my agents has a schemas_folder parameter. Will it just magically happen if I put the .json file in the schema folder, or do I need to take some steps to create a new tool? (It would be amazing if you had a GPT trained on all of the instructions!)
Looks like almost finished product. Do you plan to monetize it or provide user interface for people that have no experience in coding but need cloud access for the team? By the way, there are some powerful iPaaS platforms such as open-source active pieces. Agency Swarm could hopefully generate around 100 tools using it's code.
Hi...followed instructions from a windows terminal but jupyter notebook stalls running 4th cell when trying to import genysis module... ModuleNotFoundError: No module named 'selenium'. Any suggestions, thx.
@@IdPreferNot1Got it to run. Make sure you CD to 'notebooks' before running 'Jupyter notebook' from the command line. Had to try it a couple times. Be aware that the displayed error is cached, so you'll see it every time you load it. But pay attention to the blue indicator on the left side, showing which step is currently active. On mine, the first step is Step 2. Run them one at a time, even if the error is still displayed. Give each step time to respond. Good luck...
Hi Arsenii! I am so glad to watch videos by you on here and proud at the same time. This area is quite specific. Despite this fact, everything you talk about is so clear and it doesn't feel like smth mundane at all. I would even say, it feels like a dope thing. It's your eloquence. Surprisingly, I've been doing research on using AI as an instrument to make educational platforms more inclusive. So, I have one question (maybe, for all your subscribers on here). Tell me, do you see any ways to use AI for making businesses more socially responsible. If yes, in your opinion, can it provoke AI consciousness development? Thanks in advance XO
Can the agents see the program outputs and interact with the file system? I found that I have been spending a lot of time debugging AI written programs. It'll be wonderful if the agents can see their own outputs and correct their errors by themselves!
@vrsen really fascinated & inspired by your content. Not really experienced with coding, but decided to dive into it nevertheless. I am running into some errors though when follwoing the instructions from the video (parsing of the tools of webbrowsing agent) and the process is htoruwign error when genesis is trying to create CEA Agent. Anyone experieincing same issueS?
Can you make the gradio interface accept upload of files? I found an api that has its schema outlines in a .yaml file and the Genesis CEO was ready to try and interpret that and build the agency with that but couldn't proceed since i couldnt upload it.
This is so exciting. But can you please help me understand something? If i understood correctly, i though openAI assistant api only allowed function calling, and you hade to make these functions yourself. But can you actually make a schema, with endpoints the assistant can make get requests to? like you are doing in the video to binance endpoint?
Excellent work, downloaded and tried your project before this upgrade. But unfortunately I must say that it's absolutely useless, all these frameworks etc. that depend on OpenAI. - Which company would send their Data to OpenAI or M$, Google or Meta for processing? And now people even send them their workflows. And think about it a little more: They now also have all of the results the AI agents generated for you. - We definitely wouldn't. I am really surprised how many people would do that before thinking twice. - But maybe the privacy of data is not that important for people in the US. Yes, we are working on something like that, too, but _only_ using locally hosted models etc. - Yes, it's much more work because you first have to find the right models for every task and then eventually teach them with developing good system prompts. - But imho that's the price you have to pay for the privacy of your data. But nevertheless: Keep up the great work!
@@ryzikx They also claimed to not have trained their LLMs on copyrighted data and now it shows that they definitely lied to everybody about that. And: Even if they are not: They must be keeping the full data of your conversations with their models to be able to pick up a conversation in the past. - For how long are they keeping that data? Who can say what happens with the data when the company get's into whatever trouble or will be sold? If your company relies more and more on AI workflows and use a big company offering the AI services for that they will be able to create an image of your whole company's situation! Example: One day you give them a spreadsheet of your employees illness days of the last years to get an overview. Only with this data they will be able to say how many employees you have, how stressful working at your company is, at which points in time there was very much stress and maybe even guess when there's a bigger project to be rolled out. - If you take a bit more time thinking about what else could be found out with only these data then the 10sec I did now, you probably won't want to give those data to other companies any more. What when they get hacked? - Happened even to the best security companies themselves! - And could there be a better target than a company which has all/a lot of confidential data of other companies? Can't even imagine how much that data was worth. There's absolutely no chance we would send our or our client's company-data to a third party. And as a small or midsize company you even couldn't file a lawsuit against them because they just had to stretch it until you are bankrupt.
love the demo, trying to replicate. all good until in Jupyter step 4 (from agency_swarm.agency.genesis import GenesisAgency) getting this error: No module named 'agency_swarm.agency.genesis.AgentCreator.tools'
I clicked so quickly on this! Really love your framework and how you keep dedicating time to improve it. I hear of autogen, crewai etc, but it looks like people are not seeing how good your framework is!
Thanks, buddy!
@vrsen No, thanks to you! 🙂 You doing an amazing job!
Your framework is really amazing! If you would, future video(s) explaining how to implement your agent swarm idea "for dummies" would be really helpful to a large amount of people who dont understand more "technical" concepts.
that’s like asking Steve jobs how to code in Linux
Much rather he spends his time developing then making RUclips videos
Why not you integrate option with Langchain or other platform? So that we can easily add/change LLM model or tools
00:06 Genesys Swarm enables easy creation of AI agents without code
01:42 Agent Swarm creates Market Analyzer with new export file tool
03:12 The OpenAPI creator creates schemas from web pages for agent swarms.
04:36 Agency creation with API keys
06:02 Using various agents to analyze current Bitcoin market trends.
07:30 New updates and features in Agency Swarm
09:01 Steps to create and run Agent Swarms
10:25 The open-source AGI framework empowers community collaboration.
I would love to get a general idea of how much these swarms cost to run. Could you post a few general ranges of cost for your current tests? Nothing specific, just a ballpark idea.
Agent Swarm causing a stir with their swarms of agents SymthOS is a must-check if you're interested in cutting edge AI. It's revolutionary.
The boss is back! Less go! Always reaching further. Great work.
Thanks for support!
Extraordinary work, as always.❤, opening up a whole new dimension.
Wish i could follow your install instructions. Im that dumb with 'code' etc. But have MANY IDEAS for 'AI Swarms'. If you know of any video i should watch please suggest a basic one which would help. Excellent video and excellent 'swarms'. Well Done. Frustrated me (cant follow install instructions) but EXCELLENT WORK Dude. Thanks.
I’ll add some instruction on GitHub and post it on pip before next video
You are extremely KIND and THOUGHFUL for doing this. THANK YOU SIR!! @@vrsen
Think about a way the major agents instruction code could be autofinetuned based on real performance. Like read your prompt, read your output, how your prompt could be updated to work better, self update your prompt. Or the senior Agent does this.
Yes great idea. The challenge is in allowing agents to measure the performance of other agents. Without this it’s hard to know how to modify the prompts
It could be implemented like asking the agent that is doing the task to get his own metrics and suggest improvements for himself and then the boss to implement it.@@vrsen
Great idea, but I think the problem with limited capacities and especially a limited number of repeating user-questions is that you need to repeat those measurements a lot for the same and slightly modified questions which you can't do in a working environment because a lot of offered prompt modifications will lead to worse output, which you don't want under no circumstances in your productive setup (because you already put a lot of work into the system prompts like they are ;-).
Imho modifying finally good working system prompts are not worth of trying to optimize them more and more to gain speed.
But investigation the multiple steps (agent calls) a framework takes to solve more complex tasks and then trying to reduce the steps, that will increase the overall speed an save you a lot of money if you are using paid APIs.
And what also may speed up complex tasks is giving the agents tools that can do more than one simple thing but maybe two or more related tasks which may be called by parameters, with arrays as output.
So instead of calling multiple agents after another to each of them get one value which the next agent processes, maybe the "tool" may get all single required answers at once, give it back and only one last agent call may aggregate all that data and present the final answer.
The tools could also have a parameter with which you may say that the tool will use the gained info in a pipeline and return one value or get the answers parallel an return an array. - Using async calls in the tool when it has to generate multiple parallel answers will speed up again.
But it probably really needs a very well thought out and tested system prompt for letting a model decide if and when to use a tool in parallel or pipeline mode.
Ouuuu... it's all so challenging :-)
Great ai agent swarm guide! You are the swarm king 👑
Everything here is simply brilliant, you are a true modern scholar!
Did I just stumble upon...the Swarm King!? I came here for a CrewAI video and left with a blown mind.
Testing and working great, please add support for local LLMs too, I would love to test this framework with Mixtral, thanks for you work!
Nice work. Question. How does agency-swarm compare with and interact with other similar projects like AutoGen, ChatDev and CrewAI? Are they able to work together?
thank you it's crystal clear. furthermore what would be the difference between agency swarm and LangChain. Seams to me that there are overlaping.
Very nice, could this run with a local open source LLM of choice instead of OpenAI?
do you know if it's possible, instead of using openai api key, to use local models with litellm's api wrapper (i think that's what it has)? i will try it myself soon but just wanted to know if you had any idea
I see your platform is based on OpenAI. I hear lots of talk about using LOCAL LLMs (even customize them), but so far I couldn't find evidence that it actually works.
It could save money, allow for more privacy guarantees on local data, and make you less dependent on OpenAI's evolution (at least for some parts). Do you see any use?
Astonishing Vrsen, as always 🔥🔥
WOW that's wild looking forward to see if it will ever work with local LLMs. You are literally doing things above imaginable)
Amazing results! Given it sends this many messages, do you know how much the creation and execution were in API costs, respectively?
The most expensive is certainly the browsing agent. You can try remove it if costs are a problem
Great Job! - Consider setting up a Discord channel to build a community for easier interaction? if not, any good communities you can recommend?
I’ll think about it🤔
@vrsen thats amazing work, can these agent functions be implemented with autogen studio to use with local LLM's ?
Can we run this with llama? Or other open source LLMs?
You are so next level bro!! Keep up the excellent work and keep building, PLEASE!!
umm... is an 'Oooofff'!!!! in order here?! Wow!
Amazing.
Is this the start of Skynet? ;)
😉
you should be able to drop in lm studio endpoint so you can use opensource models.
# Chat with an intelligent assistant in your terminal
from openai import OpenAI
# Point to the local server
client = OpenAI(base_url="localhost:1234/v1", api_key="not-needed")
history = [
{"role": "system", "content": "You are an intelligent assistant. You always provide well-reasoned answers that are both correct and helpful."},
{"role": "user", "content": "Hello, introduce yourself to someone opening this program for the first time. Be concise."},
]
while True:
completion = client.chat.completions.create(
model="local-model", # this field is currently unused
messages=history,
temperature=0.7,
stream=True,
)
new_message = {"role": "assistant", "content": ""}
for chunk in completion:
if chunk.choices[0].delta.content:
print(chunk.choices[0].delta.content, end="", flush=True)
new_message["content"] += chunk.choices[0].delta.content
history.append(new_message)
# Uncomment to see chat history
# import json
# gray_color = "\033[90m"
# reset_color = "\033[0m"
# print(f"{gray_color}
{'-'*20} History dump {'-'*20}
")
# print(json.dumps(history, indent=2))
# print(f"
{'-'*55}
{reset_color}")
print()
history.append({"role": "user", "content": input("> ")})
Is the broswer taging open source too? It's been a problem I've been meaning to build a solution for generalizing navigating web broswer
I haven't been able to get the web_browser_agent to work with the genesis_agency. Am I supposed to start that first or is that a part of the genesis_agency jupyter notebook?
Hello mate! This is an amazing project! Congratulation on doing this;
I've runned it on my side and I was trying to connect my agents to Mongodb, salesforce and AWS S3 bucket but the agents can't connect to it even with the appropriate keys given to it; Do you have any reference about how can I make it works? Any help is really appreciate.
My favourite channel !
Wonderfull, but I guess if there's any method to reduce the token request speed, as i am continuously running into "OpenAI Run Failed. Error: ', 'Rate limit reached for gpt-4o."
I am also guessing if there's any way to select other models, such as gpt-3.5-turbo (apart from modifying the Python code underlying Genesis library, which I've tried to do, and btw didn't work)
Hi @vrsen, how could I increase the skills of each agent? I.e. I have one article writer + one article editor. Let's say I want to train the article editor by providing 3,000 words of examples of edited text. What's the best way to do this? Many thanks
This is extraordinary work. Congratulations. I agree with the final remarks around AGI.
I really hope that Azure can release an Assistants API equivalent!
Hello! Nice work, and I'm really exited to build my first Agency Swarm! Do you have a community where we users can help each other? I've succeeded in installing Agency Swarm, but I'm having trouble understanding how to create tools using an Open API schema. I have the schema json file. One of my agents has a schemas_folder parameter. Will it just magically happen if I put the .json file in the schema folder, or do I need to take some steps to create a new tool? (It would be amazing if you had a GPT trained on all of the instructions!)
It will magically happen if you just specify the folder) Docs are coming soon for sure, community probably as well
Hi, great work, I heard about CrewAI lately and wonder how Agency Swarm compare to CrewAI?
Looks like almost finished product. Do you plan to monetize it or provide user interface for people that have no experience in coding but need cloud access for the team? By the way, there are some powerful iPaaS platforms such as open-source active pieces. Agency Swarm could hopefully generate around 100 tools using it's code.
Hi...followed instructions from a windows terminal but jupyter notebook stalls running 4th cell when trying to import genysis module... ModuleNotFoundError: No module named 'selenium'. Any suggestions, thx.
Same...😡
@@jjgravelleReally want to check this agent system out. Does it even run on windows? WSL2?
@@IdPreferNot1Got it to run. Make sure you CD to 'notebooks' before running 'Jupyter notebook' from the command line. Had to try it a couple times.
Be aware that the displayed error is cached, so you'll see it every time you load it. But pay attention to the blue indicator on the left side, showing which step is currently active.
On mine, the first step is Step 2. Run them one at a time, even if the error is still displayed. Give each step time to respond.
Good luck...
Thx @jjgravelle. I'm running linux on WSL so i had to install a GUI browser into linux, then it opened.
Man your channel is a goldmine! so much unique stuff
Hey just found your channel, I'll likely want to contribute to this codebase my friend.
Hi, is it possible to run with Gemini Pro?
Great video! Do you do consulting work as well? I would really like to ask you some things but the calendly link is not working..
Is your framework depending on openai 100%??
Hi Arsenii! I am so glad to watch videos by you on here and proud at the same time. This area is quite specific. Despite this fact, everything you talk about is so clear and it doesn't feel like smth mundane at all. I would even say, it feels like a dope thing. It's your eloquence.
Surprisingly, I've been doing research on using AI as an instrument to make educational platforms more inclusive.
So, I have one question (maybe, for all your subscribers on here). Tell me, do you see any ways to use AI for making businesses more socially responsible. If yes, in your opinion, can it provoke AI consciousness development? Thanks in advance XO
Can the agents see the program outputs and interact with the file system? I found that I have been spending a lot of time debugging AI written programs. It'll be wonderful if the agents can see their own outputs and correct their errors by themselves!
Excited to try this tomorrow!
@vrsen really fascinated & inspired by your content. Not really experienced with coding, but decided to dive into it nevertheless. I am running into some errors though when follwoing the instructions from the video (parsing of the tools of webbrowsing agent) and the process is htoruwign error when genesis is trying to create CEA Agent. Anyone experieincing same issueS?
how can i make sure that the agency swarm only uses gpt 3.5?
Set model parameter accordingly in all of your agents
thanks! miraré el trabajo, tiene buena pinta. Se puede mezclar con agentes langchain u otros frameworks para sumar habilidades de agentes? ;)
You are the best! Congratulations!
Great work!
Can you make the gradio interface accept upload of files? I found an api that has its schema outlines in a .yaml file and the Genesis CEO was ready to try and interpret that and build the agency with that but couldn't proceed since i couldnt upload it.
This is so exciting. But can you please help me understand something? If i understood correctly, i though openAI assistant api only allowed function calling, and you hade to make these functions yourself. But can you actually make a schema, with endpoints the assistant can make get requests to? like you are doing in the video to binance endpoint?
can i use it with ollama?
Amazing. You are really on to it. More power to you.
can you make a tutorial of how to use this ... tried figuring it out ... super complicated would love a tutorial
Working on docs. Coming soon
@@vrsen 🫡
@vrsen how's docs going?
What ... The ... Fuck ...
This is insane.
Yeah, try it out haha
This is a future. I am your fans
I am having the same error message I have conflicting dependencies
You can now install from pip
Anyone get an agent running yet? I'm close but can't seal the deal.
let's go bro. so glad i found your channel
Excellent work, downloaded and tried your project before this upgrade.
But unfortunately I must say that it's absolutely useless, all these frameworks etc. that depend on OpenAI. - Which company would send their Data to OpenAI or M$, Google or Meta for processing? And now people even send them their workflows.
And think about it a little more: They now also have all of the results the AI agents generated for you.
- We definitely wouldn't.
I am really surprised how many people would do that before thinking twice. - But maybe the privacy of data is not that important for people in the US.
Yes, we are working on something like that, too, but _only_ using locally hosted models etc. - Yes, it's much more work because you first have to find the right models for every task and then eventually teach them with developing good system prompts. - But imho that's the price you have to pay for the privacy of your data.
But nevertheless: Keep up the great work!
openai claims to not train on data generated from api calls. whether its true who knows
@@ryzikx They also claimed to not have trained their LLMs on copyrighted data and now it shows that they definitely lied to everybody about that.
And: Even if they are not: They must be keeping the full data of your conversations with their models to be able to pick up a conversation in the past. - For how long are they keeping that data?
Who can say what happens with the data when the company get's into whatever trouble or will be sold?
If your company relies more and more on AI workflows and use a big company offering the AI services for that they will be able to create an image of your whole company's situation!
Example: One day you give them a spreadsheet of your employees illness days of the last years to get an overview.
Only with this data they will be able to say how many employees you have, how stressful working at your company is, at which points in time there was very much stress and maybe even guess when there's a bigger project to be rolled out. - If you take a bit more time thinking about what else could be found out with only these data then the 10sec I did now, you probably won't want to give those data to other companies any more.
What when they get hacked? - Happened even to the best security companies themselves! - And could there be a better target than a company which has all/a lot of confidential data of other companies? Can't even imagine how much that data was worth.
There's absolutely no chance we would send our or our client's company-data to a third party.
And as a small or midsize company you even couldn't file a lawsuit against them because they just had to stretch it until you are bankrupt.
Best & Useful Content ! Thanks a lot Dude. Keep up the good work. I have a request, can you try to build an open source version of perplexity ai.
Thank you!
Thanks for watching!
Uses the openAI assistants API??? It’s gonna be really expensive and really slow lol
Vrsen is the best :)
Super cool!
@vrsen. In a word...."Awesome" 🎉
Thank you
hows it differnt then new autogen hmmm
Thats Dope! Kudos Bro
Amazing!
are all the agents you are using defaulting to gpt-4-preview model its super expensive , i dont think this will work at scale
You can change this by changing model parameter when initializing an agent
🤯
This looks insanely good…any plans to drop in something like openrouter or flowise/langchain…that openai API is Uber expensive for now….
amazing content. keep it coming
awesome
Replacing all boring low level human and engineering jobs. Next level!
Love it!
Great stuff! APIs are terrible, I found that Claude can create better quality api connectors in Go than in python
your channel just took off. agents are exploding. ;-)
love the demo, trying to replicate.
all good until in Jupyter step 4 (from agency_swarm.agency.genesis import GenesisAgency)
getting this error: No module named 'agency_swarm.agency.genesis.AgentCreator.tools'
cancel. figured it out! ready to play