Pricing
Start with a free account. Talk to us for Enterprise.
We don’t publish prices. A free account, by invitation, lets your team try Sigma. Enterprise covers deployment on your own infrastructure or in a sovereign cloud region, and is quoted in a proposal for your organisation.
Free
For teams that want to try Sigma on a real use case.
Free, by invitation
- Invitation-only: request access and we review every request
- Build agents from instructions, skills and tools
- OpenAI or Anthropic models, chosen per agent
- Limits shared on invitation
Enterprise
For organisations that need to run Sigma in their own environment.
Quoted for your deployment
- Deployment on-premise, on your own infrastructure, or in a sovereign cloud region you choose
- Support for your security review: architecture and data-flow documentation, the third-party services your deployment uses, and answers to your questionnaires
- Onboarding for your team
- Terms and pricing in a proposal scoped to your deployment and scale
Today, every deployment sends prompts to the model provider you choose (OpenAI or Anthropic) and document chunks to a managed vector index. In-country inference and fully offline deployment are on the roadmap. See Trust and security.
Pricing FAQ
Questions about pricing and accounts
Why don’t you publish prices?
Enterprise cost depends on how you deploy (on-premise or a sovereign cloud region), how many people and agents you plan for, and the onboarding you need. So we quote it in a proposal rather than a price list.
How do I get a free account?
Accounts are invitation-only. Fill in the request access form with a few details about your organisation and how you want to use Sigma. We review every request and reply by email.
What are the free account’s limits?
We share the limits that apply with your invitation.
Can we move from a free account to Enterprise?
Yes. Contact us when you are ready to plan a deployment on your own infrastructure or in a sovereign cloud region.
Can Sigma run on-premise or in a sovereign cloud?
Yes, on Enterprise. The platform runs on your infrastructure or in your cloud account on Microsoft Azure, AWS or Google Cloud. Today it needs outbound access to your chosen model provider and to the managed vector index used for document search. In-country inference and fully offline deployment are on the roadmap.
Which AI models can we use?
OpenAI or Anthropic models today, chosen per agent. Other providers and self-hosted or in-region models are on the roadmap.
Not sure which fits?
Tell us what you want agents to do and where they need to run. We’ll suggest the right starting point.