内容来源：https://blog.n8n.io/top-up-your-n8n-assistant-credits/

内容总结：

n8n云服务今日起支持Assistant额度充值。Starter和Pro版用户在月度额度用尽后，可直接购买额外额度继续使用，无需再等待每月1日额度重置或被迫升级套餐。

充值方式

用户可选择手动充值或开启自动充值。自动充值可设置触发余额和补充目标，并默认启用月度消费上限。系统会优先使用套餐包含的额度，用完后才扣除额外额度。额外额度可跨月结转，有效期为购买后12个月。

额度使用情况可视化

云管理后台新增Assistant额度标签页，可查看包含额度与额外额度的余额明细、近24小时至30天的消费图表、实例内用量最高的用户以及充值记录。编辑器中的Assistant面板会显示当前对话已消耗额度，并在剩余10%时发出提醒。

月度额度已更新

Starter和Pro用户的月度额度有所调整。官方表示，对大多数用户而言新额度更耐用，原因包括：迁移至更新、更强的模型，完成相同任务所需额度更少；额度价值在所有付费套餐中实现统一；底层优化使每次请求效率更高。

典型任务额度消耗参考

额度消耗因任务复杂度而异。简单提问约10至20额度；单次提示完成简单工作流或修复约50至100额度；包含测试的大型构建或较长调试会话约300至600额度。官方提示，提示词越具体，消耗的额度越少。

注意事项

额度在实例内共享，所有成员共用同一额度池。仅实例所有者可充值，且需有效付费套餐。运行工作流不消耗Assistant额度。Assistant额度与Gateway额度相互独立，分别用于对话和AI模型调用。免费试用提供800额度，试用期间不可充值。自托管版n8n不使用Assistant额度，Assistant通过用户自有模型提供商账户运行。

中文翻译：

从今天起，如果你使用的是 n8n Cloud（Starter 或 Pro 版），并且用完了 Assistant 额度，你可以购买更多额度，继续构建。

自从我们推出 n8n Assistant 以来，能够充值一直是我们收到的最常见的需求之一。在此之前，额度用完意味着你只能等到每月 1 日额度重置，或者升级你的套餐。当你正构建工作流构建到一半时，这两个选择都不太理想。

充值如何运作

你的套餐仍然包含每月一定额度的 Assistant 额度，在每月 1 日重置。充值是在此之外的额外额度，并且是可选的。

你可以通过两种方式添加额外额度：

• 手动充值。 需要时随时选择额度包。
• 开启自动充值。 选择触发充值的额外额度余额阈值，以及要充值到的余额。月度上限默认开启，限制自动充值每月可扣款的金额。

Assistant 始终优先使用你套餐包含的额度，只有在这些额度用完后才会使用额外额度。额外额度可以逐月结转，并在购买后 12 个月过期。

查看你的额度去向

Cloud 管理后台新增的 Assistant 额度标签页会显示：

• 你的余额，分为包含额度和额外额度
• 过去 24 小时、7 天或 30 天的消费图表
• 你实例上哪些用户使用的额度最多
• 你的充值历史记录

在编辑器中，Assistant 面板会显示当前对话已使用的额度，并在额度剩余 10% 时提醒你。

我们已更新你的每月额度

如果你使用的是 Starter 或 Pro 版，你会注意到每月额度已发生变化。

对大多数人来说，更新后的额度会比之前更耐用。这得益于我们做出的一些改进：

• 更新、更强大的模型。 我们已将所有人迁移到更新的模型，因此 Assistant 完成同样的工作需要更少的额度。
• 统一的额度价值。 既然你可以购买额度，额度就需要有一个固定的价格。我们重新校准了额度，使其在每个付费套餐中价值相同，并相应调整了每个套餐的额度。
• 底层优化。 我们优化了 Assistant 的内部运作方式，使每次请求能完成更多工作。

构建一个工作流需要多少额度？

这是我们最常被问到的问题，但没有统一的答案。这取决于请求的内容：更长的对话、更复杂的工作流和调试过程比简短、有针对性的请求消耗更多额度。由于 LLM 行为的非确定性以及不同的上下文环境，没有两个提示的成本会完全相同。

我们能提供的是一个大致的范围预期。以下是目前几种典型请求的额度消耗：

• 一个简单问题 [10-20 额度]
• “n8n 支持 Slack 的哪些认证方式？”
• 通过一个提示构建简单工作流或修复问题 [50-100 额度]
• “每天一次，使用免费的天气 API 查询柏林的天气，并将预报通过邮件发送给我。”
• 需要明确测试的较大构建，或较长的调试过程 [300-600 额度]
• “在构建之前，先研究当前关于免费无需密钥 API 的文档，包括地理编码、天气预报、国家信息和汇率，并挑选三个本周天气差异很大的真实目的地，使每个目的地落入不同的旅行类别。然后构建一个‘旅行天气简报’工作流：表单询问城市、旅行日期和本国货币，上述 API 收集数据，撰写打包建议，将旅行标记为炎热、寒冷、多雨或温和，每份简报保存到数据表中。将简报发布到 Slack 频道，并在表单的完成页面上展示。确保处理未知城市或货币等错误，并用你研究的三个城市测试该工作流。”

这些范围会随着我们持续改进 Assistant 及其背后的模型而变化，因此请将它们视为粗略参考。一个肯定有帮助的建议：具体的提示比开放式的提示消耗更少的额度。

几点需要了解的事项

• 额度在你的实例内共享。 实例上的所有人共用同一个余额。如果你团队中有五个人使用 Assistant，他们使用的都是同一个额度池，用量最多用户视图会显示谁在使用多少。
• 只有实例所有者可以充值。 充值还需要一个有效的付费套餐。
• 你的工作流会继续运行。 运行工作流永远不会消耗 Assistant 额度，即使该工作流是由 Assistant 构建的。只有你与 Assistant 的对话才会消耗额度。
• Assistant 额度和 Gateway 额度是分开的。 Assistant 额度用于支付你与 Assistant 的对话。Gateway 额度用于支付你的工作流和代理在运行时发起的 AI 模型调用。充值其中一个不会增加另一个。
• 免费试用包含一次性额度。 免费试用提供 800 额度。试用期间无法充值，因此你需要升级到付费套餐才能获得更多额度。
• 自托管的 n8n 不使用 Assistant 额度。 Assistant 改为在你的自有模型提供商账户上运行。

如需了解所有详细信息，请查看 Assistant 额度文档。

有问题？

直接在社区论坛帖子中回复即可。

英文来源：

Starting today, if you're on n8n Cloud (Starter or Pro) and you run out of Assistant credits, you can buy more and keep building.
Being able to top up has been one of the most common requests we've heard since we launched n8n Assistant. Until now, running out meant waiting for your allowance to reset on the 1st of the month, or upgrading your plan. Neither is a great option when you're halfway through building a workflow.
How top-ups work
Your plan still includes a monthly allowance of Assistant credits that resets on the 1st of the month. Top-ups are extra credits on top of that, and they're optional.
You can add extra credits in two ways:

• Top up manually. Pick a credit package whenever you need one.
• Turn on auto top-up. Choose the extra-credit balance that triggers a top-up and the balance to refill to. A monthly limit, on by default, caps how much auto top-up can charge in a month.
  The Assistant always uses your included credits first, and only moves on to extra credits once those run out. Extra credits carry over from month to month and expire 12 months after you buy them.
  See where your credits go
  The new Assistant credits tab in the Cloud admin dashboard shows:
• Your balance, split into included and extra credits
• A spend chart for the last 24 hours, 7 days, or 30 days
• Which users on your instance use the most credits
• Your top-up history
  In the editor, the Assistant panel shows how many credits the current conversation has used, and warns you when you're down to 10%.
  We've updated your monthly credits
  If you're on Starter or Pro, you'll notice that your monthly credit allowance has changed.
  For most people, the updated allowance will go further than before. That comes down to some improvements we’ve made:
• A newer, more capable model. We've moved everyone to a newer model, so the Assistant needs fewer credits to do the same work.
• One standard credit value. Now that you can buy credits, a credit needs a fixed price. We recalibrated credits so they're worth the same on every paid plan, and set each plan's allowance to match.
• Improvements under the hood. We've optimized how the Assistant works internally, so it gets more done with each request.
  How many credits does it take to build a workflow?
  That's the question we get most often, and there's no single answer. It depends on the request: longer conversations, more complex workflows, and debugging sessions use more credits than short, focused requests. No two prompts will cost the same, due to non-deterministic LLM behavior and different surrounding context.
  What we can provide is a is a sense of the range to expect. Here are what a few typical requests use today:
• A quick question [10-20 Credits]
• "What authentication methods for Slack does n8n Support?"
• "What authentication methods for Slack does n8n Support?"
• A simple workflow or fix from one prompt [50-100 credits]
• "Once a day, check the weather in Berlin using a free weather API and email me the forecast."
• "Once a day, check the weather in Berlin using a free weather API and email me the forecast."
• A larger build with explicit testing, or a longer debugging session [300-600 credits]
• “Before building, research the current docs for free keyless APIs for geocoding, weather forecasts, country info and exchange rates, and pick three real destinations with very different weather this week so each one lands in a different trip category. Then build a "Travel Weather Briefing" workflow: a form asks for a city, travel dates and home currency, the APIs above gather the data, writes packing tips, labels the trip Hot, Cold, Wet or Mild, and each briefing is saved to a Data Table. Post the briefing to a Slack channel, show it on the form's completion page. Ensure you handle errors such as unknown cities or currencies, and test the workflow with the three cities you researched.”
• “Before building, research the current docs for free keyless APIs for geocoding, weather forecasts, country info and exchange rates, and pick three real destinations with very different weather this week so each one lands in a different trip category. Then build a "Travel Weather Briefing" workflow: a form asks for a city, travel dates and home currency, the APIs above gather the data, writes packing tips, labels the trip Hot, Cold, Wet or Mild, and each briefing is saved to a Data Table. Post the briefing to a Slack channel, show it on the form's completion page. Ensure you handle errors such as unknown cities or currencies, and test the workflow with the three cities you researched.”
  These ranges will change as we keep improving the Assistant and the models behind it, so treat them as a rough guide. One tip that definitely helps: specific prompts use fewer credits than open-ended ones.
  A few things to know
• Credits are shared across your instance. Everyone on the instance draws from the same balance. If five people on your team use the Assistant, they're all using the same pool, and the top users view shows who's using what.
• Only the instance owner can top up. Topping up also requires an active paid plan.
• Your workflows keep running. Running a workflow never uses Assistant credits, even if the Assistant built it. Only your conversations with the Assistant do.
• Assistant credits and Gateway credits are separate. Assistant credits pay for your conversations with the Assistant. Gateway credits pay for the AI model calls your workflows and agents make when they run. Topping up one doesn't add to the other.
• Free trials include a one-time allowance. Free trials get 800 credits. Top-ups aren't available during a trial, so you’ll need to upgrade to a paid plan to get more.
• Self-hosted n8n doesn't use Assistant credits. The Assistant runs on your own model provider account instead.
  For all the details, check out the Assistant credits docs.
  Got any questions?
  Just send a reply in the community forum post.