Usage Limits

eustella Usage Limits & Fair Use

Last updated: 06.07.2026

This page explains how usage limits work on eustella so you know what to expect before you subscribe. It is provided for transparency.

This page is about how much you can use eustella. For the rules on how you may use it — what is and isn't allowed — see the Usage Policy.

One shared weekly allowance

eustella meters usage as a single, shared weekly allowance. Chats, image generation, web search, file reading, and agent runs all draw from the same pool, so you can spend your allowance however suits you rather than being locked into separate per-feature quotas.

As a rough guide, one Free plan is enough for about 200 chats per month.*

These figures are indicative equivalents of the same allowance, not independent quotas — using one draws down the other, because they share one pool.

Paid plans give you the same allowance, multiplied — so you can do more of either, in any mix:

See Pricing for what each plan includes. Want to get more out of every chat? Find out how to maximise your usage in our best practices guide.

* A note on these figures. The chat and image counts above are a gross approximation — not the actual number of chats possible, which varies a lot by use case. eustella calculated them with our best practices in mind, without any so-called tool calls (web searches, document creation, memory writes, and the like) and with a basic Soul setting for eustella. Real usage that involves tool calls, longer conversations, or a richer Soul draws down your allowance faster, so treat these numbers as a rough orientation rather than a promise.

How the allowance works

eustella meters your usage with these elements:

• Weekly allowance — your main allowance. This is the allowance described above. It resets each week.
• 5-hour rolling limit — a short-term guard (Star and Cosmos). On the Star and Cosmos plans, a shorter rolling limit measured over the last 5 hours sits alongside your weekly allowance, so one intense burst can't consume it in a single sitting. Other plans do not have this limit.
• Top-up credit — an optional overflow pool. If you add credit, it is used only after your regular allowance is full — never before — and it carries over rather than resetting.

If you reach your limit, some capabilities pause until your allowance resets; your data and history are unaffected. In the app, your usage is shown as a percentage of your current allowance rather than a running unit count — the figures above are indicative equivalents of that same allowance. Some accounts are unmetered (no limit); those accounts see no usage bar.

What consumes your allowance — and how much

Everything eustella does runs on the underlying AI model, and your allowance is what pays for it. Two kinds of activity draw it down.

Thinking and writing. Every reply means eustella reasoning through your request and writing an answer — this is the baseline cost of any chat message. A short question costs little; long, involved answers and long conversations cost more, simply because there is more for the model to read and to produce.

Tool calls. To act rather than only talk, eustella uses tools — searching the web, reading a file or a memory, looking at an image, generating a picture, and more. Depending on the tool, a call either costs about what it takes to read the information it returns (that text enters the chat like any other input), or it costs more when the tool itself is expensive to run. In detail:

• Reading something costs about what pasting it would. When eustella reads a memory, opens a file, or looks at an image you send, that content becomes part of what the model has to take in — so it costs roughly what it would if you had copy-pasted that text or document into the chat yourself. A short note is cheap; a long report costs more. Nothing hidden — simply the length of what is read.
• Web search costs on two counts. First, the European provider we use for the search index charges us a little for every search. Second, the results read back count as ordinary input — exactly as if that text had arrived in your chat. One quick lookup is minor; a session full of searches adds up.
• Image generation costs more. Creating an image means calling a larger, more expensive image-generation model, so a generated picture uses noticeably more of your allowance than a written reply. Worth knowing before you produce a dozen variations.
• Agents cost what the same work would cost by hand. An agent draws from the same weekly allowance as your chats: every search it makes, every document it reads, every image it creates, and the report it writes all count, just as they would in a normal chat. The one thing to watch is that a schedule repeats — a daily research agent quietly spends a full research session's worth of allowance every day, whether or not you read the result. So schedule deliberately: choose the cadence you will actually use, not the most frequent one available. If an agent reaches your limit, its run simply pauses and retries in the next window, and your schedule stays intact. The 5-hour guard on the Star and Cosmos plans (above) still applies.

By contrast, a plain chat message is about the cheapest thing you can do — another reason short, focused chats stretch your allowance further.

More allowance means more freedom — not perfect answers

If you subscribe to a paid plan, first of all: thank you for supporting us. A subscription buys you more usage allowance — a larger shared pool to spend across chats, searches, images, and agents, and so more room to do more with eustella.

What more allowance does not buy is perfect, error-free answers. Like every AI today, eustella can still make mistakes or hallucinate on any plan — a bigger allowance gives you more room to work, search & run agents, but it is not a guarantee of accuracy. For how to get the best and most reliable results from eustella, see our guide: How to use eustella — best practices.

Business and API use

For business and API customers, usage limits, rate limits, and quotas are set out in your plan or in an individual written agreement, as described in the Business Terms. API rate limits are documented separately for developers.

Changes to allowances

eustella may adjust these allowances over time as models, costs, and usage patterns evolve.

In the future

• More powerful model options. eustella may introduce higher-capability model options. If it does, those models consume more of your allowance per action than the standard model, so the same allowance stretches to fewer heavy-model actions. The equivalents above are estimated at the standard model.

Questions

If you have a question about your allowance or believe a limit has been applied in error, contact us at support@eustella.com.

© 2026 AI Newsrooms Technology GmbH. All rights reserved.