Engagements and pricing
Private AI deployments vary by workload, model or runtime, infrastructure, integrations, controls, availability, and operating responsibility. Izwi reviews those requirements before providing pricing.
Proposal inputs
Engagement options
Start by evaluating the runtime, defining the architecture, proving one workflow, deploying to production, or adding ongoing operations.
Product evaluation
Best for: Developers validating private speech workflows locally.
How it works: Download the public runtime and test it on your own machine. Contact us when you need production packaging, deployment guidance, or support.
Download Izwi →Assessment
Best for: Teams that need a documented architecture and production recommendation.
How it works: Tell us about the workload and environment. We’ll return a proposal with the questions to answer, the work involved, and the deliverables you’ll receive.
Request assessment pricing →Fixed-scope pilot
Best for: Teams ready to test one real workload in one target environment.
How it works: We price the pilot against agreed success criteria, the target environment, infrastructure responsibilities, and test requirements.
Request pilot pricing →Implementation
Best for: Production private voice or LLM infrastructure.
How it works: Your proposal covers the implementation, integrations, responsibilities, delivery plan, and production requirements for your environment.
Request deployment pricing →Managed service
Best for: Deployments needing ongoing monitoring, maintenance, upgrades, and support.
How it works: We price ongoing service around the deployed systems, coverage hours, availability needs, change cadence, and division of responsibilities.
Request managed-operations pricing →Proposal factors
Pricing questions
Tell us about the workload and target environment. We’ll return a proposal covering Izwi’s work and identify infrastructure, GPU capacity, third-party software, and model-license costs separately where relevant.
Tell us which service you need, what the system must do, where it will run, and your timeline. We’ll review the requirements and recommend the right engagement.
The work changes substantially with the workload, infrastructure, integrations, scale, availability, and support requirements. Pricing a defined system avoids hiding those differences in an arbitrary tier.
Usually, your organisation owns the hardware or cloud capacity directly. We identify infrastructure, model licences, and other third-party costs separately so the proposal is clear.
Izwi's core commercial model covers software, assessment, engineering, deployment, support, and operations. Any workload-based component is stated explicitly in a customer proposal.
The Voice AI Runtime supports local evaluation. Work involving real infrastructure, security boundaries, and production criteria normally begins with an assessment or paid pilot.
Share the workload, target environment, infrastructure responsibility, integrations, security requirements, and support model.