Powered by
The intelligence layer that scores every task and routes it to the best model in real time. Learn more →
You choose from providers supported by your deployment. ToggleLogic applies configured routing policy and records its decisions; model selection remains deployment-dependent and subject to review.
The intelligence layer that scores every task and routes it to the best model in real time. Learn more →
You always work with SAM. The resources behind SAM have different roles, and your deployment determines which ones are enabled.
The accountable hub: your continuing business context, memory, policies, approvals, and customer-facing conversation.
Reasoning engines selected under ToggleLogic policy for quality, cost, speed, availability, and privacy requirements.
Defined capabilities that let SAM work with approved files, applications, services, and business processes.
Bounded workers for assignments such as software development, research, analysis, or content production. They are optional and remain subordinate to SAM's workflow.
You do not have to manage multiple assistants. SAM coordinates enabled resources and returns the work through one relationship, subject to the deployment's permissions and owner gates.
You choose which models SAM connects to. Today, SAM ships with support for six models across four providers — but if it has an API, SAM can use it. Add a new provider, remove one you don't need, or swap in the latest model the week it launches. ToggleLogic handles the routing. You set the rules.
Your installed language model runs continuously on your Mac mini. No API calls. No token costs. No data leaving your building. The workhorse that handles the majority of daily tasks before any cloud model is ever called.
Top-tier reasoning for complex, nuanced tasks. When SAM needs to think deeply — strategy, writing, code, analysis — Sonnet is the call.
Fast, accurate, and a fraction of the cost of Sonnet. Ideal for background automation and tasks that need reliability without deep reasoning.
Ultra-fast and cost-effective. Excels at multimodal tasks — images, documents, and structured data — at speeds that keep SAM responsive.
Deep general knowledge and conversational fluency. Strong at creative tasks, summarization, and scenarios requiring broad world knowledge.
Real-time awareness and unfiltered reasoning. Fast, capable, and built for tasks that benefit from current information and direct responses.
If it has an API, SAM can connect to it. Add new providers as they launch, swap models as pricing shifts, or bring your own fine-tuned model. ToggleLogic routes them all.
SAM isn't a generic chatbot loaded with everything. The skills SAM runs depend on who you are and what you do. Every SAM deployment starts with a common foundation — then adds profession-specific capabilities tailored to your goals.
The core capabilities that every business needs — regardless of industry. These run out of the box on day one.
SAM learns what matters to you. A franchise operator gets different skills than a real estate agent or a financial advisor. Your AI agent works the way your business works.
Dynamic Model Router
ToggleLogic is SAM's model-switching intelligence layer. It scores every incoming task on four dimensions and routes to the best model — not the most expensive one.
Simple file search? Local LLM. Routine classification? Haiku or Gemini. Deep strategic analysis? Sonnet or OpenAI. ToggleLogic scores complexity before any model is called.
Connected models have different price points. ToggleLogic evaluates configured cost and task requirements to reduce avoidable premium-model usage while preserving the deployment's quality policy.
Real-time response? Gemini, Grok, or the Local LLM. Can it wait? Use the most cost-efficient option. You set the priority. ToggleLogic enforces it.
You define which work must remain local and which authorized workflows may use connected cloud services. Those boundaries are enforced through the deployment's configured routing and approval controls.
The intelligence layer keeps learning. As models evolve and pricing shifts, ToggleLogic re-evaluates routing automatically. You stay optimized without lifting a finger. Learn more at ToggleLogic.ai →
Other AI platforms ask you to trust a cloud with your credentials and your memory. ToggleLogic keeps both locked to hardware you own — with patent-pending architecture that survived a real-world breach.
Every API key, every OAuth token, every credential SAM uses to access your business systems is bound to the physical Secure Enclave of your Mac mini. Nothing is ever stored on disk. Nothing is ever stored in the cloud. If the software is fully compromised, the attacker gets nothing — because there's nothing to get.
Most AI assistants either forget everything between sessions — wasting tokens reloading context — or remember everything unchecked, accumulating errors over time. Toggle Logic Memory solves both problems with a dual-tier architecture where the AI observes freely but only you decide what becomes permanent knowledge.
Hardware-Lock™ and Toggle Logic Memory™ are patent-pending technologies developed by Motherboard, Inc. and available exclusively through ToggleLogic. Read the origin story →
Every model you connect has its own price point. Local LLM is free. Lightweight cloud models cost pennies per task. Premium reasoning models are reserved for work that justifies the cost. ToggleLogic tracks every dollar across every provider you use.
See exactly what you're spending on each model, per task, per day. No surprises on your API bill.
With multiple configured models available, ToggleLogic is designed to reduce unnecessary premium-model usage while maintaining the deployment's task-quality requirements. Results vary by workload and provider pricing.
Set monthly limits per model or overall. SAM won't exceed them. Ever. Cost governance is built into the system.
Most AI platforms chain you to one provider. If that provider raises prices, degrades quality, or goes down — you're stuck. SAM runs on ToggleLogic, which means switching models is a configuration change, not a rewrite. New model launches? We add them. Provider gets too expensive? We route around them. Your intelligence layer adapts — your workflow doesn't break.
Your models. Your skills. One platform. Zero lock-in.
Get Started