请问HN：AI驱动应用的请求限制与令牌限制？
目前，我正在开发一个用于管理文档、数据库和白板的网络应用程序——这是一款典型的应用，旨在像 Notion 一样。<p>然而，现在我面临着制定一个有 AI 使用限制的计划的困境，因为我的想法是让它更具自主性：能够在整个工作区内编辑和查询上下文，并将其转移到文档中，例如，可能在白板上绘制一些东西等。不过，我感觉消费可能会很快失控。我计划使用 DeepSeek 进行 AI 聊天，但使用 Gemini 3 Flash 进行自主使用和编辑，因为它更智能。最近，我注意到许多核心 AI 应用程序已经将定价模式从按请求计费转变为固定使用限制，但我不确定这是否会受到批评，是否会导致用户体验不佳，或者甚至让人觉得没有得到所支付的价值。因此，我希望听听大家对我应该做出什么决策的看法。
查看原文
Currently, I’m working on a web app for managing documents, databases, and whiteboards—the typical app that aims to be like Notion.<p>However, right now I’m facing the dilemma of creating a plan with AI usage limits, since the idea is to make it more agentic: able to edit and query context across an entire workspace and move it into a document, for example, maybe draw something on a whiteboard, etc. Still, I have the feeling that consumption could easily get out of hand. I plan to use DeepSeek for AI chat, but use Gemini 3 Flash for agentic usage and editing because it’s more intelligent. Lately, I’ve seen how many core AI apps have shifted their pricing models from per-request pricing to a fixed usage limit, but I’m unsure whether that’s frowned upon, creates a less pleasant user experience, or maybe even gives the feeling that you’re not getting what you paid for. So I’d like to hear some opinions on what decision I should make.