STU, quotas and spending
Shara measures consumption in a single unit, the STU (Shara Token Units), independent of the model used underneath. Think of it as the fuel your agents burn: the longer or more complex a task, the more STU it uses. Your plan includes a quota, a monthly allowance of STU. If you use more than that, the extra is called overage and costs a flat rate. If you use less, part of what is left carries over to the next month (rollover). And automatic spending limits stop anything that runs away. The idea is that you never get a surprise invoice at the end of the month.
What an STU is
Section titled “What an STU is”The STU is the client’s single unit of consumption, independent of the model used underneath. The rule is direct: 1 STU = 1 weighted token. Every time an agent reads or writes, Shara counts the real tokens (input plus output), applies the weight of the model that handled the request and that is the STU. There is no divisor or hidden scale in between.
The weights do two things. Output weighs more than input, because it costs more to generate. And more capable models weigh more per token, so the same sentence resolved by Symphony spends more STU than if resolved by Prelude. A model’s weight is proportional to what it really costs, so every STU costs the same on any model and spending the whole quota on the most expensive one doesn’t give you less work than you paid for. Input the provider serves from its cache counts for a quarter.
The conversion is deterministic: for a given model and a given volume of tokens, the STU cost is always the same. There are no hidden rates or opaque rounding.
| Tier | Alias | Weight per token |
|---|---|---|
| Agile | Prelude | The lowest of the three cloud models: input 0.5, output 1.25 |
| Production | Sonata | Input 3.125, output 11.25 |
| High capacity | Symphony | The highest: input 4.375, output 11.25 |
| Local | Concerto | 0 · runs on your hardware, doesn’t use quota |
In round numbers: a task of 8,000 input tokens and 1,200 output tokens comes to about 5,500 STU with Prelude, about 38,500 with Sonata and about 48,500 with Symphony. It gives you the order of magnitude, not a rate card.
Since Amadeus delegates each task to the most suitable model, your consumption naturally goes down: routine tasks fall to Prelude or Sonata, and only complex reasoning reaches Symphony. You never pay for more capacity than a simple task needs.
How consumption works in each interaction
Section titled “How consumption works in each interaction”Every exchange with an agent deducts STU from your monthly quota. An interaction’s consumption depends on three things:
- Input size: the message, plus the context the agent needs (its
IDENTITY.md, the relevant memory, the email or document it’s reading). - Output size: what the agent drafts or reasons through to respond.
- The model handling it: the higher the tier (Prelude, Sonata, Symphony), the more STU each token weighs.
A task can involve several chained calls. If Amadeus receives a request and delegates it to a departmental agent, and that agent in turn queries a tool and drafts a document, each step adds its part to the interaction’s total. The dashboard breaks that journey down so you can see exactly what was spent on what.
How far it goes in practice
Section titled “How far it goes in practice”Quotas are sized for real business use, not to be burned through in an afternoon. As an indicative reference (the exact consumption depends on the model, the context and the length of each response):
- A short task (classifying an email, extracting data from a message, drafting a brief reply) is low consumption, especially when handled by Prelude or Sonata.
- A medium task (drafting a proposal, summarizing a long thread, preparing a report with data) consumes more, because the context and the output both grow.
- A complex task (analyzing a contract clause by clause, a strategic analysis, in-depth research) weighs the most, and is usually resolved by Symphony.
Translated into monthly capacity, each plan covers on the order of thousands of interactions. This is a sizing guide, not an exact promise: what’s reliable is looking at your real consumption on the dashboard during the first few weeks and adjusting the plan from there.
| Plan | STU / month | Usage profile it covers |
|---|---|---|
| First | 50M | An essential team automating its day-to-day |
| Pro | 140M | Intensive daily use with more departments |
| Max | 280M | Intensive multi-department operation |
| Max ×5 | 1,150M | Large teams with high concurrency |
| Max ×20 | 4,000M | High volume for a mid-sized company |
Quotas by plan
Section titled “Quotas by plan”| Plan | Price | STU / month |
|---|---|---|
| First | €129/month | 50M |
| Pro | €349/month | 140M |
| Max | €690/month | 280M |
| Max ×5 | €2,900/month | 1,150M |
| Max ×20 | €9,900/month | 4,000M |
| Concerto Local | Contact sales | 60M in the cloud (optional hybrid component) |
| Additional agent | +€99/month | +15M |
Six-month billing is charged as five months out of six and annual as nine out of twelve, both with a commitment for that period. Concerto Local (custom pricing, through sales) deploys Shara on your own infrastructure, everything included. It also comes with 60M STU/month for the Amadeus component in the cloud of the hybrid mode, which you switch on or off yourself: switched off it consumes nothing, and it never generates overage. Full detail of each plan in Plans and pricing.
Overage and quota levels
Section titled “Overage and quota levels”If you exceed your monthly quota, work doesn’t stop at 100%: consumption moves into overage, billed at a flat rate. Overage accumulates during the period and is added to your next invoice.
| Plan | Overage |
|---|---|
| All paid plans | €5/M STU |
| Concerto Local | No overage |
The rate is flat and identical across every paid plan: €5 for each million STU you spend above your quota. Concerto Local doesn’t apply: its cloud pool is fixed and never generates overage.
The percentage is measured over your quota for the period (base plus rollover). Here is what happens at each level:
| Consumption | Level | What happens |
|---|---|---|
| 0 to 75% | green |
Normal consumption. |
| 75 to 90% | warning |
Warning: you’ve used a significant share of your quota. |
| 90 to 100% | critical |
Few STU left. Time to consider changing plan. |
| 100 to 130% | overage |
Overage is billed at €5/M. |
| 130 to 200% | grace |
Overage plus forced saving mode (see below). |
| 200% or more | hardstop |
Inference is blocked: the API answers 429 quota_exceeded until the period resets. |
So the service isn’t cut at 100%. From 130% Shara moves to saving mode (cheaper models, see below), and at 200% of your quota it stops until the next period.
Rollover: the quota you don’t spend
Section titled “Rollover: the quota you don’t spend”Rollover rewards steady use by carrying over to the next month the STU you didn’t consume. It only applies to cloud plans (Concerto Local’s pool is fixed and doesn’t roll over) and works like this:
- The rollover is what you left unused of your base quota:
rollover = base quota − consumed, with a floor of 0. It is calculated against the base quota, not against the rollover you already carried. - It has a cap of 20% of the base quota. Even if you have more left over, the next month you carry over at most that 20%.
- There is a floor of 1%. If in the month you consume less than 1% of your base quota, nothing is carried over: the system treats the plan as not in use. And if you consume 100% of the base quota or more, there is nothing left to carry over.
- It doesn’t accumulate. The next month’s rollover replaces the previous one, it isn’t added to it.
- It expires after 2 months. Rollover not used within that window is lost.
The rollover is added to the month’s quota: overage only kicks in once base plus rollover are used up. This way, a slow month doesn’t penalize the next one.
Example with Pro (140M STU/month): if in a given month you consume 120M you have 20M left over, which is below the 28M cap (20% of 140M), so you carry over the full 20M. If you had consumed only 70M, 70M would be left over, but the cap applies and you would carry over 28M. The following month you start with that rollover on top of the base quota.
Consumption alerts
Section titled “Consumption alerts”So you’re never caught by surprise, Shara alerts you as you approach your quota limit. You choose at which percentages of consumption you want to be warned among 60%, 80%, 90%, 100% and 150%. If you never choose, the defaults are 80%, 90% and 100%. Each threshold warns at most once per billing period; if a single jump crosses several, you get one alert for the highest.
- 80% and 90% (by default): proximity-to-limit alerts. Worth deciding whether to activate saving measures or plan for overage.
- 100%: you’ve used up the period’s quota. From here consumption moves into overage (€5/M).
- 150%: well into overage.
- 60% (optional): an early alert, to check whether the month is going to be a heavy one.
Separately, the 200% cut (hardstop, see above) always notifies, whatever alerts you have chosen. Alerts reach the account administrator and stay visible on the dashboard, along with the consumption breakdown by agent and by day.
Saving mode
Section titled “Saving mode”Outside the business hours you set in Settings → Business hours, Shara automatically forces low consumption. The idea is that the work you launch at night or on weekends (with the Run button or by asking Amadeus) keeps running without driving up cost.
- Automatic model downgrade: Symphony drops to Sonata and Sonata drops to Prelude, so the same task consumes less STU.
- Output limit on Prelude: capped at 2,500 tokens for more contained replies outside business hours.
- No service cut: whatever you launch or ask Amadeus for keeps working; only spending is reduced.
Saving mode is also forced when your consumption is between 130% and 200% of your quota (grace level). Concerto Local never drops to a lower model.
In practice, saving mode noticeably cuts costs outside business hours without stopping you from working. Business hours are set for your company, so it adapts to your real working day.
Spending limits: protection against runaway spending
Section titled “Spending limits: protection against runaway spending”Spending limits are automatic cuts. They work on three levels, and they are set in STU, not in euros. Their purpose is to stop a loop or a badly configured integration before it burns through your quota.
| Level | Default limit | What it does |
|---|---|---|
| Per run | 1M STU in a single execution | Stops the run and pauses the company’s inference. |
| Per hour | 10M STU in the last hour | Pauses inference and alerts the admin. |
| Per day | 80M STU in the last 24 hours | Pauses inference while the window stays above the limit. |
The owner or an admin can edit each limit or turn it off from the app. The hour and day windows are rolling, not clock-based: they look at the last 60 minutes and the last 24 hours from right now. As a guide only, at the overage rate those limits are worth €5, €50 and €400.
When a limit trips, inference stops with a notice and the cut is recorded in the history on the Usage screen. The hourly and daily cuts lift on their own once their window drops below the limit, and the per-run cut after one hour; an admin can lift them earlier, and can also set a manual pause that doesn’t lift on its own. In Concerto Local they don’t apply in local-only mode: there is no cloud cost to measure.
How to control spending
Section titled “How to control spending”Between the quota, the rollover, the alerts, saving mode and the spending limits you have several levers to keep consumption under control. In practical order:
- Size the plan to your real consumption. Look at the dashboard the first few weeks and move up or down a plan based on the STU you actually spend, not on estimates.
- Set your business hours correctly so saving mode covers the hours when you don’t need the premium models.
- Choose your consumption alerts (60%, 80%, 90%, 100%, 150%) and watch them to decide in time whether you expect overage or need to adjust use.
- Take advantage of rollover: what you leave unused, up to 20% of the base quota, carries over to the next month for 2 months.
- Keep the spending limits active and adjust them to your tolerance; they’re the safety net for an unexpected spike. If one trips, something has run away and it’s worth a look.
- Use overage as a buffer, not as the norm. At €5/M it’s predictable for occasional peaks; if it’s recurring, it pays off to move up a plan.
To choose a plan and see full pricing, go to Plans and pricing. The billing engine measures every call by its real cost and applies your plan’s margin, with no opaque flat rates.