Taskade has grown in every direction at once. It started as a collaborative task and outline tool, added documents, then AI agents, then automations, and in October 2025 launched Genesis, a prompt to app builder. The help centre has kept pace admirably in the narrow sense that all of it is documented, four hundred and eleven articles across ten collections, and if you want to know how a particular button behaves you will find out. As reference material it is competent and well maintained.
As education it barely exists, and the difference matters because the product now spans so much ground that a user's real question is not how does this button work but what am I supposed to be doing here. There is a Learn Taskade hub alongside the help centre carrying several hundred more articles, and adding volume does not add a path. There is no sequence, no progression, no worked scenario that carries a reader from a real problem to a built solution. You get an index of capabilities and are left to assemble a purpose from it.
For a tool whose whole pitch is that it consolidates several categories of software, refusing to explain how those pieces fit together is the central failure. The AI agent material is the clearest instance. It explains how to create an agent, give it instructions, connect it to a workspace and let it act. What it never addresses is when you should not.
An enormous proportion of the workflows people try to automate with agents would be better served by a template, a filter, a scheduled reminder or a two line rule, and would then be predictable and free and would not quietly do the wrong thing at three in the morning. Guidance that helped users tell those cases apart would produce far better outcomes than more feature documentation, and no vendor writes it because it means talking somebody out of using the expensive part of the product. Genesis has the same problem in sharper form. Prompt to app is the current gold rush and the documentation treats it as a capability to be operated rather than a practice to be learned.
Building software by describing it has real failure modes around specification, iteration, understanding what was generated and maintaining it afterwards, and none of that is discussed. Pricing is at least clear, with Free, Pro at ten dollars a month, Business at twenty five, Max at a hundred and Enterprise at two hundred and fifty, all billed annually, and it is genuinely reasonable. Nothing here is bad. It is simply the wrong artefact for the situation, a filing cabinet where the product needed a teacher.
Two point nine. Thorough, current, free reference documentation, marked down because a product this sprawling needed someone to explain the point of it and nobody did.