Simple, transparent pricing.

No hidden fees. No surprise bills. Everything is priced in EUR.

Start with a defined 90-day pilot. We connect Freshdesk, ingest your knowledge base, and you see cited draft replies within 14 days.

Team

Scale across your support team

For teams past the pilot: more projects, more team members, and full audit logs.

€750 /month
50 GPU-hours / month GPU Talk to us
  • Unlimited projects & assistants
  • Unlimited documents
  • Team members & permissions
  • Multiple scoped API keys
  • Usage & audit logs
  • Webhook support
  • Priority email support
  • DPA available on request
  • GDPR erasure and export
  • EU-hosted data
  • 50 GPU-hours / month included (advanced fine-tuning)
Advanced

Enterprise

Advanced: fine-tuning, dedicated infrastructure

For teams that need SSO, dedicated deployment, custom model providers, or unlimited fine-tuning compute.

Custom
Unlimited GPU Contact us
  • Unlimited projects & assistants
  • Data Processing Agreement included
  • Custom data retention agreement
  • SSO / SAML — Planned Q1 2027
  • SLA available on request
  • Private or dedicated deployment
  • Custom data residency
  • Custom model providers
  • Unlimited GPU training time (LoRA / QLoRA fine-tuning)
  • Connect your own GPU servers
  • Dedicated onboarding & support
  • Custom billing terms

Advanced: GPU fine-tuning add-on (Team and Enterprise)

Optional, for teams past the pilot with specialised domain language. Buy extra GPU-hours when your monthly quota is exhausted. Set a spending limit to cap usage.

€2 / GPU-hour

All Cortex plans include

Cited answers

Every draft links back to the exact document and section used.

Human review before send

Cortex drafts. Your team reviews and sends. Never the other way round.

Audit log included

Every generated answer, and who accessed what, is logged.

EU-hosted by default

Hetzner infrastructure, EU data centres. GDPR-ready out of the box.

End-to-end isolation

Your data is isolated at the project level. No cross-customer access.

OpenAI-compatible API

Drop Cortex into any stack that already talks to OpenAI. No code changes.

Frequently asked questions.

What exactly do I get in the 90-day support pilot?

A Freshdesk integration (Symphony), your docs and old tickets ingested into a knowledge base, and reviewable AI draft replies with citations. We promise a working pilot within 14 days.

What do you measure during the pilot?

Search time per ticket, first response time, draft acceptance rate, citation accuracy, escalation reduction, and agent onboarding time - tracked on your own tickets, not a demo.

Does Cortex send replies without a human?

No. Every draft is queued for a support agent to review, edit if needed, and send. Cortex does not send automatically.

Do I need fine-tuning to start?

No. The Support Pilot and Team plans use retrieval and drafting only. Fine-tuning is an advanced, optional capability on the Team and Enterprise plans for teams with specialised domain language.

How does GPU training time work on the Team and Enterprise plans?

The Team plan includes 50 GPU-hours per month for optional fine-tuning. Extra GPU time is available at €2/GPU-hour with a configurable monthly spending limit. Enterprise plans have no quota and can connect their own GPU servers.

How does Canopy Host work?

Canopy Host is a separate, secondary product. Push your code to a connected Git repository, Canopy detects your stack, builds a container, and deploys it. Custom domains, SSL, and environment variables are managed from the dashboard.

Is my data used to train shared models?

No. Your documents, conversations, and training data are processed solely to power your own assistants. We do not use your data to train any shared or platform-level models.

Where is data hosted?

By default, all data is stored on EU-based infrastructure (Hetzner, EU region). Customers requiring US or APAC residency should contact us before signing up.

Can I cancel at any time?

Cortex plans are month-to-month after the pilot term unless agreed otherwise. Canopy Host Starter is free with no credit card required.

Can I connect my own model provider?

Yes. Cortex supports Anthropic, OpenAI, and other compatible providers. You bring your own API keys, which are encrypted at rest.

Not sure which plan fits? Let's talk.

hi@canopy.pm