Activity
Mon
Wed
Fri
Sun
Jan
Feb
Mar
Apr
May
Jun
Jul
Aug
Sep
Oct
Nov
What is this?
Less
More

Memberships

AI Developer Accelerator

Public β€’ 3.8k β€’ Free

22 contributions to AI Developer Accelerator
Deploying CrewAI Apps to some hosting??
Hi guys. I'm a medium developer, but I can mostly get things to work the way I need in the end. However, the bane of my existence is the "production" side of things. Installing dependencies and all the stuff that makes my mind numb with dullness. Which begs the question: Does any one have a good guide or suggestion for deploying a crew to some place where it can be reached as an API (and no, CrewAI's on setup did nothing for me). πŸ™
1
7
New comment 7d ago
0 likes β€’ 8d
PS. That course looks aaaaamazing @Brandon Hancock !! I can't wait to dig in. What a great idea!! πŸ’ͺ🏻πŸ’ͺ🏻πŸ’ͺ🏻πŸ’ͺ🏻
0 likes β€’ 7d
@Tom Welsh Thanks mate! I went to bed last night with only one last issue. When I'm calling my endpoint to run the crew I get a 405 error 🀯 I'll post a video and some details about it later today. I think I'll need some help from the community. But I think in general it would be a nice topic for the forum to address. As a hobby- or fresh developer (which I suspect a lot of us are), ideas and coding is what we're spending time on. The DevOps bit is a lot more engineer heavy. Providers like Vercel and Render.com has made it super easy to deploy a web site or sweb app, but getting a Docker image online requires a whole different skillset. Anyways, thanks for help and support πŸ™πŸ»
New Fullstack AI Tutorial With CrewAI Enterprise
πŸ‘‹ Hey everyone! If you’ve been itching to build a full-stack AI application with real-world impact, this is the tutorial you’ve been waiting for. In my latest video, you’ll learn how to create an app that connects directly to CrewAI Enterprise, the platform that makes deploying and monitoring your crews incredibly easy. πŸ”₯ Who Is This For? - Beginner-to-intermediate developers. - Anyone looking to turn their Crews into a SaaS app. πŸ› οΈ What You’ll Use: - πŸ–₯️ NextJS, Drizzle, Neon, and TypeScript for full-stack development. - πŸ€– CrewAI Enterprise for AI-powered decision-making and automation. 🎯 What You’ll Build: A full-stack AI app that helps creators figure out their next video ideas: - πŸ“ Data Scraping: Pull videos and comments from a YouTube channel. - πŸŽ₯ AI Processing: Use CrewAI to generate video ideas, titles, descriptions, research trends, and score ideas. - βš™οΈ Crew Integration: Seamlessly deploy and monitor AI workflows with CrewAI Enterprise. πŸ“š What You’ll Learn (3 Phases): 1️⃣ Phase 1: Project Setup - Lay the foundation with modern web technologies. 2️⃣ Phase 2: Code and Deploy the NextJS App - Build the app and take it live. 3️⃣ Phase 3: Build and Deploy Your Crew - Harness the power of CrewAI to handle AI-driven tasks. πŸ’‘ Why Watch This? - πŸ› οΈ Get hands-on experience building a full-stack AI app. - πŸš€ Learn how to integrate and optimize workflows with CrewAI Enterprise. - πŸ“ˆ Add modern tools to your developer toolkit. ✨ Bonus: I’m giving away the source code for free so you can follow along and experiment with your own projects! πŸ’ͺ Want to Level Up Even Further? If you’re interested in turning this app into a SaaS product, check out the AI Developer Accelerator Pro. πŸ‘‰ I’ve included a mini-module on scaling this project to automate weekly YouTube scraping and idea generation. πŸ”— Learn more here: https://www.skool.com/ai-developer-accelerator-pro/about
17
12
New comment 7d ago
New Fullstack AI Tutorial With CrewAI Enterprise
1 like β€’ 8d
totally stoked πŸ”₯
New Claude Sonnet Model and Cursor
The Cursor community have updated the models to include the new claude-3-3-sonnet-20241022 model. My current project is still using the older model. Looking on the Cursor forum, there are currently a few reports that the newer model is not behaving as well as the old model. https://forum.cursor.com/t/claude-3-5-sonnet-updated-and-claude-3-5-haiku-soon/24165/11 I might leave it a bit before jumping in a new project. That said. I am intrigued but the 'computer use' part of the updated 3.5 Sonnet model.
3
4
New comment 22d ago
2 likes β€’ Oct 22
How do you all think this will affect agents? In the examples, people are typing and the process is slugish (no shade). However, we already know that a conversation with a computer has near perfect fidelity already (like with OpenAI's realtime API). I'm giving it 6 months before I can talk to my computer to make it do things on screen. What then?
1 like β€’ 24d
@Idael Pineiro Yeah, I think we should just assume that future models will handle a computer as well as us (most probably better) for long tasks. Maybe we'll see it in 6 months πŸ€·πŸ»β€β™‚οΈ Maybe 5 years. Who knows. My money is on sooner rather than later. But then I start thinking.. ok the internet today is built for humans. If we build AI that can use the internet/computers better and faster than us. Will the 'human web' become obsolete? If so.. won't we just create a new internett/OS that is more ideal for machines? Do we then go back to the human interfaces? But what would be the point?
Recurring tasks in Flows?
Has anyone tried to set up a recurring flow? Like a job that keeps running like every 15 mins after last completion? Did a short Loom on the subject to get some opinions: https://www.loom.com/share/a5396b76ca484585876ad70834dc81e1?sid=b55baf34-d062-4ed3-a227-4b49e8c81634 I guess my main thing is: - Should I try to do it all in one flow? If nothing else than just for the hell of it? - Can I get it to just loop?
0
1
New comment 29d ago
0 likes β€’ 29d
Ok. After some experimenting it seems like creating a loop in the actual flow is not the way to go. However, it wasn't very hard to use apscheduler to run the flow every 4 hours. I think its the easiest to just let the flow go through creating the recipe first, and then using the state to write a post on some similar topic. And the once the flow is done, just wait for a new run. We can probably use the state to set some global variables to guide the state for the next run as well. Any other insights or ideas are more than welcome
Traning crews in Flows
Hi guys. I have a flow that will use 3 crews (probably more). However, the boilerplate has no 'train' and 'test'. Has anyone tried to create a new crew as a "standalone" crew, train and optimize it, and then just drop it into a flow?
0
0
1-10 of 22
Michael Birkeland
3
39points to level up
@michael-birkeland-6703
Hobby developer and super AI enthusiast 😍 Norwegian.

Active 6d ago
Joined Sep 6, 2024
powered by