Cloud inference is billed separately from local licensing.
Your quote specifies the models, token rates and billing terms for your deployment.
Components and supported hardware are agreed for each deployment. Customization and ongoing services are priced separately by contract.
Customization and services
From models to deployment.
By contract
Adapt individual components or develop an integrated Personal Intelligence Layer for your product.
Custom language models
Specialize PLM for your domain, terminology and workflows, including reasoning, structured outputs and tool use.
Custom voice models
Customize KRORAINA 2’s voice, pronunciation and expressive delivery using authorized voice data. Tune interaction behavior for your product.
Hardware-specific runtime
Optimize the Inference Framework for your target hardware, with quantization, scheduling and memory use tuned to its compute and latency constraints.
Hardware integration and tuning
Integrate the models and runtime with your device’s audio, applications and tools. Tune and validate the complete system on the target hardware.
Post-training
Build training data from domain examples and approved feedback. Apply supervised fine-tuning and preference optimization, and evaluate gains alongside general capabilities and safety.
Ongoing support
Technical support, troubleshooting and maintenance for deployed systems, including model and runtime updates under your support agreement.
Model, voice and runtime changes are evaluated together on the deployment configuration.
Yes. The Personal Intelligence Layer combines PLM, KRORAINA 2, the Inference Framework, memory, interaction control and Smart Router. We can provide an integrated system or adapt individual components for your product. The deployment agreement specifies the components and supported hardware.
Can we customize one component without replacing the others?
Yes. Language, voice and interaction control can be customized separately. We test changes in the complete system to verify their effect on quality, timing and resource use.
How do we get a quote?
Tell us about your product, deployment environment and expected device count or usage volume. Include any customization, integration or support requirements so we can prepare a proposal.
What is the difference between deployment and activation pricing?
Both are one-time license fees. A deployment license covers the system and deployment scope specified in your contract. Activation licensing charges a fee for each device when it is activated.
Is there an inference charge on our hardware?
There is no metered SADAIR inference fee for work completed on your hardware. Hardware and operating costs, and any use of remote models or tools, are separate.
Can local and cloud deployment work together?
Yes. Local models can handle the live interaction while selected tasks use remote models or services. Third-party model and tool charges are not automatically included in SADAIR’s cloud pricing.
How are customization, maintenance and updates priced?
Customization, maintenance and updates are priced by contract. Your agreement defines deliverables, maintenance coverage, update terms and support, separately from local licensing or cloud usage.
