I hate giving this product such a low review, mainly because Sean is exceptional and genuinely dedicated to making this work. However, there's no getting around that the company prioritizes finances over product quality.

I purchased the highest AppSumo tier and the multi-agent system add-on for about $1,000 expecting a functional multi-agent platform. Five months later, I've used roughly 100 credits monthly from my 200,000 allocation but not from lack of engagement. I'm on the platform every other day, watch every office hour, and read just about every Discord post. I know the status on the platform, from a user perspective. If you're someone who likes to build and troubleshoot agents for fun, this platform will give you that. But if you like to use agents that are already built and reliable, this is not the platform for you. Users spend more time learning what terms/coding mean vs working with their agent.

Sean is genuinely brilliant with an impressive vision. But good intentions don't resolve fundamental stability issues. Watch any recent office hour and you'll witness restarting agent chats multiple times or switching to advanced models when agents fail mid-demo. Even the in-house copilot doesn't work reliably, forcing users to external platforms (Claude/ChatGPT) just to generate workflows/etc for agents that still may not function correctly due to the platform bugs. This platform isn't like (Sintra, Marbalism, etc)

Instead of stabilizing their platform, they showcase other companies' features (Claude, ChatGPT, Google releases) and discuss integrating them into an already unstable system. Their current architecture will not allow agents to perform tasks consistently.

What concerns me most is their BYOK/PixelML strategy. If I need a ChatGPT subscription ($20/month) for BYOK, why pay additional monthly fees for Agentic Flow's agents once ChatGPT has already started deploying their own native agents in-house? Mix & matches LLMs with agents would be answer if it were included in the plan but it's not in ways you'd expect. The "pay as you go" method reminds me of when cellphones were pay per minute. You'd think that would feel better but users didn't like not knowing what the bill would be at the end of the month. Same applies here. User's would rather pay one price to have everything (even with limits) then having the surprise bill at the end of the month. While this deal looks "okay" from a lifetime perspective, if it lives, the subscription side falls flat, which will determine the life of this product. User are better off waiting for the AI giants to release multi-agents in-house, which literally just happened a few weeks ago.

Previous AI platforms on appsumo (Nily AI, Greta, Qolaba, etc) include advanced LLMs (Nily AI alone has Sonnet 4, Gemini 2.5 pro, veo 3, etc included in the sumo plan) because they understood users wouldn't abandon subscriptions they already have for an unproven platform. I understand that these platforms weren't also offering Agentic agents but at the very least they knew including buzzwords that users were already familiar with and pairing them with the platform subscription felt like we were getting a deal. We could also see how monthly subscriptions could be excited about the deal, given all that is included. I currently have lifetime deals with platforms like GlobalGPT and Galaxy Ai, Nily (both are still offering lifetime deals), that offer agents & almost every available LLM (image, video, chat) that has been made, limited only by monthly credits. Both also have subscribers who are paying because the users want to see one monthly fee vs multiple. Agentic Flow offers DeepSeek, GPT-4o mini, GPT oss 120b and other lower tier models in this plan that aren't enough to entice users.

BYOK and PixelML signal that Agentic Flow may not believe they can financially support these models long term in a plan , as they don't believe they will have the monthly subscribers to recoup those cost. To show real commitment, they'll need to take that risk and provide models that make the product shine, not expose its underdeveloped foundation.

Here's the current state...
"Appropriate" models fail due to platform bugs
Architecture requires advanced LLMs for 70-80% success rates
Their competitive window is shrinking as users won't pay two subs once AI giants provide the same services, with stability, natively.
When I detailed these issues, they claimed included models work "just fine," contradicting documented Discord experiences and showing disconnect from the customer experience.

Right now, you're paying to beta test unfinished software while being directed to additional subscriptions. As much as it pains me to write this, knowing it may hurt my chances of recouping my investment, I must be honest. Until they show real investment in themselves & the users beyond words and promises, the platform remains in beta.

Sean is a gem who truly means well, but action is key here.