A visual AI orchestrator where agents check their own work | Wazear

Hello Product Hunt, I am Mohamed, the creator of Wazear. It was created from a frustration I had. Every time I wanted to work on a project with the implementer being AI it would just end up being a total mess. Either because I end up lost between tons of documents and prompts moving them between different agents or whatever the agent has created is not critically and properly reviewed. So I built Wazear to solve this specific problem. Wazear allows you to create agents using .md files -- it already ships with a sample pipeline if you don't want to start from scratch. The pipeline resembles that of the software development cycle. You tell Wazear what you want to build, it spawns a planning agent to create a plan, a plan-reviewing agent to review that plan. They run in a loop till the plan satisfies your criteria based on the brief. Next, the architecture agent and its reviewing agent run in a loop and so on till you finally hit the QA wall where you're prompted to test yourself and take screenshots of the product and provide feedback on how to improve the final result. However, at any point in time you can simply pause the pipeline, inspect and sway the results to a different direction if you want. Wazear is 100% local (except of course for the agent instance), it doesn't collect data, it is not a SaaS, pay once, own current version forever.