Transparent pricing, no surprises
Clear monthly STU quotas (Shara Token Units). Go over your quota and the service does not stop: you pay the overage at your plan’s rate.
Plans and prices
1 month free when you pay six up front and 3 when you pay the year, which is the one that saves most · on Concerto Local, which starts at six months, only the year is discounted
First
Start with an essential AI team.
overage 5 €/M STU
- 50 M STU/month included
- Amadeus, the director, plus 5 departmental agents
- Carlzon, Carnegie, Kotler, Graham and Maslow
- Universal connectors (Gmail, Slack, Calendar, CRM, ERP, Teams)
- Desktop app (Windows/Mac/Linux) plus server agent
- IDENTITY.md per agent: they learn your tone and your processes
- Per-company isolation plus Postgres RLS
- Self-serve, running in 24 hours
- Email support (48 working hours)
- Overage 5 €/M STU
Pro
A complete AI team for your company.
overage 5 €/M STU
- 140 M STU/month included
- Amadeus, the director, plus 10 departmental agents
- Everything in First plus Deming, Holmes, Juran, Rosling and Turing
- IDENTITY.md per agent: they learn your tone and your processes
- Semantic memory across conversations
- Gmail, Calendar, Slack, HubSpot and Notion connectors
- Priority support
- Overage 5 €/M STU
Max
Executive suite, with strategy and BI.
overage 5 €/M STU
- 280 M STU/month included
- Amadeus, the director, plus 15 departmental agents
- Everything in Pro plus Porter, Brunel, Moore, Bernays and Schneier
- Top-tier model driving Amadeus
- 5 extra agents available (99 €/month + 15 M STU each)
- Capacity selector ×1 / ×5 / ×20
- Optional white-label plus a signable DPA
- Contractual SLA, signable · priority support
- Assisted onboarding · overage 5 €/M STU
+ extra agents 99 €/month (+15 M STU each)
Concerto Local
Total sovereignty: you decide what leaves.
999 €/month · 5,994 € per half-year
your data on your own infrastructure · cloud quota optional
- Full deployment on your own hardware or private cloud
- Everything included: the app, the local Concerto engine and every agent
- The Concerto models run on your server: that work never leaves your network
- Three tiers to match your infrastructure: Concerto Compact (small servers), Concerto Standard (mid-range servers), Concerto Extended (large servers, maximum throughput)
- 20 agents included: the 15 from Max plus the 5 add-ons, at no extra cost
- 60 M STU/month of cloud quota, optional: you switch it on or off · no international data transfer
- Ed25519 licensing with a 30-day grace period
- Single-tenant on your own infrastructure · encrypted with your own KMS
- Dedicated support in European business hours plus a monthly incident drill
- Built for regulated sectors
A 6-month minimum: Concerto Local is installed on your own infrastructure, on site or remotely, and that setup does not pay for itself in a single month. Six months is already the shortest term, so the half-yearly discount does not apply here; the annual one does.
Need more capacity than Max ×20? Talk to us about a bespoke enterprise plan. Prices in EUR, VAT included.
- 4
- Plans
- First · Pro · Max · Concerto Local
- 21
- Specialist agents
- 5 €
- Overage per M STU
- A flat rate on every cloud plan
- EU
- Hosting and inference
- Zero international transfers
At no extra cost
Included on every plan
From First up. What changes between plans is the number of agents, the model behind Amadeus, the STU quota and the support.
Amadeus, the orchestrator, on every plan
Universal connectors: desktop app, server agent, email, chat, calendar, CRM and ERP
Approval queue (human in the loop): critical actions wait for your sign-off
Automatic low-spending mode outside working hours
Spend kill-switches: 5 € per run · 50 € per hour · 400 € per day
Baseline security: per-company isolation, AES-256 encryption, MFA and anti-prompt-injection
Data hosted in the EU (Concerto Local: on your own infrastructure)
You own your data outright: export to JSON/CSV
Three periods: one month, six months or a year. Month to month, no lock-in on the cloud plans
Comparison
What each plan includes
The Max ×5 and Max ×20 variants scale capacity (STU) on top of the same functional Max base.
| Feature | First | Pro Recommended | Max ×1 · ×5 · ×20 | Concerto Local |
|---|---|---|---|---|
| Model capacity | ||||
| The director’s model (Amadeus) Capacity grows with the plan | Standard | Advanced | Top-tier | Top-tier, local |
| Departmental agents’ model | Standard | Standard | Advanced | Local, on-premise |
| Catalogue and capacity | ||||
| Amadeus, the director | Included | Included | Included | Included |
| Departmental agents | 5 (Carlzon · Carnegie · Kotler · Graham · Maslow) | 10 (+Deming · Holmes · Rosling · Turing · Juran) | 15 (+Porter · Brunel · Moore · Bernays · Schneier) | 20 (the same as Max plus the 5 add-ons, on-premise) |
| Add-on agents Barnum · Curie · Shannon · Tracy · Ohno | No | No | 5 available (99 €/month each) | 5 included, at no extra cost |
| Monthly STU quota (base) On Concerto Local the cloud quota is optional: you switch it on or off, and what runs on your own server spends no STU | 50 M | 140 M | 280 M | 60 M in the cloud · no quota on your server |
| Capacity scaling The ×5 and ×20 variants scale capacity on the same functionality | No | No | Max ×5 (1,150 M) · Max ×20 (4,000 M) | Your own hardware |
| STU rollover Use 80 % or more and up to 20 % carries into the next month; it expires after 2 months | Yes | Yes | Yes | No |
| What the agents can do | ||||
| IDENTITY.md per agent Captures your tone, vocabulary, FAQs and processes | Yes | Yes | Yes | Yes |
| Semantic memory | No | Yes | Yes | Yes |
| Approval queue (human in the loop) Critical actions (sending email, making a payment, hiring) require approval | Yes | Yes | Yes | Yes |
| Automatic low-spending mode Outside working hours it drops to a cheaper model | Yes | Yes | Yes | Yes |
| Automatic kill-switches 5 € per run · 50 € per hour · 400 € per day, cut off automatically with a notice to the admin | Yes | Yes | Yes | Yes |
| Continuous learning (IDENTITY patches) | No | Yes | Yes | Yes |
| Connectors and integrations | ||||
| Email (Gmail / Outlook) | Yes | Yes | Yes | Yes |
| Chat (Slack / Teams) | Yes | Yes | Yes | Yes |
| Calendar (Google / Outlook) | Yes | Yes | Yes | Yes |
| CRM (HubSpot · Pipedrive) | Yes | Yes | Yes | Yes |
| CRM (Salesforce) On the roadmap | Coming soon | Coming soon | Coming soon | Coming soon |
| ERP (Holded, Spain) | Yes | Yes | Yes | Yes |
| ERP (Odoo) On the roadmap | Coming soon | Coming soon | Coming soon | Coming soon |
| Notion | Yes | Yes | Yes | Yes |
| Airtable · Shopify · and more We add integrations every month | Coming soon | Coming soon | Coming soon | Coming soon |
| Custom MCP connectors Model Context Protocol, your own extensions | No | No | Yes | Yes |
| Desktop app (Win/Mac/Linux) plus server agent | Yes | Yes | Yes | Yes |
| Security and compliance | ||||
| Per-company isolation (RLS plus middleware) Two layers: Postgres RLS and the tenantContext middleware | Yes | Yes | Yes | Single-tenant on your infrastructure |
| Encryption at rest (AES-256-GCM envelope) | Yes | Yes | Yes | ✓ (with your KMS) |
| Encryption in transit (TLS 1.3) | Yes | Yes | Yes | Yes |
| MFA (TOTP) | Yes | Yes | Yes | Yes |
| Append-only audit log Every run is written down and cannot be edited or deleted | Yes | Yes | Yes | Yes |
| Anti-prompt-injection (red team) 54 attacks from the OWASP LLM Top 10 plus Shara-specific ones · 95 % or more neutralised | Yes | Yes | Yes | Yes |
| Where data is hosted | EU | EU | EU | Your infrastructure |
| International transfers Inference processed entirely inside the EU | None | None | None | None |
| Signable DPA plus full audit logs | No | No | Yes | Yes |
| White-label / your own branding | No | No | Optional | Yes |
| Operations and support | ||||
| Onboarding | Self-serve | Self-serve plus an optional demo | Guided by a consultant | Pilot plus a consultant |
| Support channel | Email plus Slack/chat | Email, chat, Telegram and phone | Dedicated, EU business hours | |
| Support By email, with a person behind it, in European working hours | Yes | Priority | Priority | Priority |
| Availability SLA | No | Best effort | Contractual SLA, signable | No cloud SLA · monthly drill |
| Monthly incident drill | No | No | No | Yes |
| Commercial | ||||
| Billing periods One month, six months or a year. The period you choose is charged up front | Monthly · 6 months · annual | Monthly · 6 months · annual | Monthly · 6 months · annual | 6 months · annual |
| Free months for paying up front The half-year is 16% off and the year 25%: committing to a year pays off. On Concerto Local the half-year is the list price, because it is its minimum term | 1 when you pay 6 · 3 when you pay the year | 1 when you pay 6 · 3 when you pay the year | 1 when you pay 6 · 3 when you pay the year | 3 when you pay the year |
| Trial period Pilots on Max and Local are agreed case by case | No | No | Guided demo | Pilot |
| Minimum term Whatever period you pick: paying month to month, none. Concerto Local is installed on your own infrastructure, and that setup does not pay for itself in a month | Month to month: none | Month to month: none | Month to month: none | Six months minimum |
| Getting your data back | JSON/CSV, 30 days | JSON/CSV, 30 days | JSON/CSV, 30 days | Backup plus the repo on your infrastructure |
| Sales cycle | Self-serve | Self-serve plus a demo | Guided sale | Consultative sale |
Consumption and cost governance
Overages, add-on agents and kill-switches
Go over your quota and the service keeps running, you add capacity when you need it, and spend always has a ceiling.
STU overages
If you go over your quota, the service is not interrupted:
- First5 €/M STU
- Pro5 €/M STU
- Max and above5 €/M STU
Rollover at 80 % consumption or more, capped at 20 %, expiring after 2 months. Cloud plans only.
Add-on agents (Max)
On Max you can switch on extra specialists (Barnum, Curie, Shannon, Tracy, Ohno):
plus 15 M extra STU per agent
Switched on from inside the app, with no contract to renegotiate.
Kill-switches
Automatic spend cut-offs:
- Per run5 €
- Per hour50 €
- Per day400 €
If an agent tries to go past the threshold, Shara stops the run and notifies the admin.
Not sure which plan to choose?
Tell us your task volume and we will recommend the right plan within 24 hours. No commitment.
Ready for your AI team?
We show you Amadeus live, with the connectors and agents you actually need. A 30-minute demo built around you.