Limits and spending

What limits use and spending, who sets them, and what happens at one?

Nearly every limit is a ceiling set per module that admins can only lower. AI use is paid through the module's own provider key, but connected apps think on their own accounts. Keep in mind: A connected app's requests to the module's bots and its searches by meaning still use the module's own provider, which also summarizes and indexes files the app writes when those features are on. One bot setting can be raised, the optional per-person token cap and the module's optional storage cap are set by admins rather than shipped as ceilings, and some ceilings have no Settings control yet. A model server you run yourself never receives the key, and AIEUS's built-in reading of text from scanned PDFs and images uses no provider, unless an admin has selected Google Document AI for it.

Limits are ceilings on how much an AI reply, file, search or import can use. Almost every one ships at the highest value AIEUS allows, so it can be turned down but not up. Keep in mind: Raising a ceiling takes a new AIEUS release. A few ceilings have no setting yet, so they cannot be turned down either. There is one exception, for bots (AI assistants inside AIEUS). The limit on how many messages one read of a bot conversation returns ships at 500 and can be raised to 2,000.

Almost every limit is set per module. A module is a self-contained space in AIEUS with its own files and people. Only a module's admins can open and change its limit settings. A nested module does not inherit its parent's choices. Keep in mind: Members cannot open these settings, but the values are not hidden from them. The limits apply to everyone in the module, and a bot reply that stops at a limit names that limit.

Partial: A bot stops a reply early at a reply limit and names that limit. By default, a reply may use up to 256 actions, one hour and one million tokens (units of text AI models count). Keep in mind: Token counts are approximate. A scheduled reply gets at most five minutes. When one bot asks another, the answer must fit in the time left. This was checked against the code, not yet on a live installation. AI apps you connect to AIEUS yourself are not bound by these limits.

AIEUS places only one limit on a person, rather than on one reply, file or request. It is an optional daily cap on the tokens used by the bot replies that person asks for. An admin can switch it on. Keep in mind: The cap is off by default. It counts each person's last 24 hours on a rolling basis. It is checked only when a new reply starts, so a reply already running can finish past it. AIEUS has no usage page.

Through the Web App, AIEUS's browser interface, you can upload files or whole folders. Each file can be up to 5 gigabytes by default. Originals such as PDFs and images are stored unchanged. Keep in mind: AI agents and connected apps can create, read and edit only text files, up to 32 megabytes each by default. Larger or non-text files must be uploaded through the Web App.

A text file's contents become searchable up to 32 megabytes by default. A larger file is still kept, but its text cannot be searched. Keep in mind: Search catches up a moment after each change, so a file you just created may not show up right away. Wait a moment or open it directly. Searching by meaning works only after an AI provider is selected for file search. An AI provider is the service that supplies the AI model.

AIEUS itself imposes no pricing tiers or plan limits. It ships with no storage cap, but an admin may set one for a module. Any other plan-style limit comes from your hosting arrangement or AI provider account. Keep in mind: A storage cap applies only once it is set. After that, saves that would go over it are refused, but files already stored stay. A nested module does not inherit the cap.

There are no prices, plans, free tier or service-level agreement.