Questions Answered

361 is an AI Operating System designed for enterprise businesses. With 17 AI providers, 80+ built-in models (400+ via OpenRouter), and 15 ready agent templates, it injects intelligence into every layer of your business.

It doesn't replace your current system — it empowers it. Backed by 20+ years of enterprise experience with 200+ modules across 14+ sectors.

No. 361's core philosophy is "Don't Replace, Empower." It installs alongside your existing business system, reads your data, and adds an AI layer.

It creates value without touching your infrastructure. If you don't like it, nothing has changed.

The free POC process takes 2 weeks. Week one: your current system is analyzed and AI scenarios are identified. Week two: pilot implementation.

Choose from 6 entry scenarios: Natural Language Reporting, Email Automation, Anomaly Detection, Document AI, Forecasting, or Workflow Automation.

Ready-made AI Solutions templates install in minutes. Installation starts with one click; the solution skeleton is ready within minutes.

Custom solutions typically take 2-4 weeks. The POC process is completed in 2 weeks.

We support 17 AI providers: OpenAI, Claude, Gemini, OpenRouter, Groq, Ollama, MiniMax, NVIDIA, Mistral, DeepSeek, Together, Cohere, Qwen, LM Studio, 361 Local (embedded llama.cpp), OpenCode, and OpenCode Go.

Automatically select the best model-cost combination for each task. Never be locked into a single provider.

REST API, webhooks, IMAP email, database connections, Microsoft 365, Google Workspace integrations.

10 communication channels: WhatsApp, Teams, Slack, Telegram, Email, Discord, Web Chat, SMS, LINE, and Voice AI.

Yes. Multi-layered protection (4 detectors + guardrail pipeline + audit trail): PII masking (national IDs and bank accounts auto-hidden), hallucination control (Truth Engine), prompt injection protection, cost limiting, and GDPR/KVKK compliance.

All AI interactions are logged. Data localization, deletion rights, and consent management are supported.

We serve 14+ sectors: Finance, Manufacturing, Logistics, Retail, Energy, Construction, Hospitality, Automotive, Healthcare, Education, Telecommunications, Insurance, Real Estate, and Sustainability.

Sector-specific AI Solutions templates are available for each. Best practices from 300+ projects.

Pricing is based on usage and scale. Start with a free POC — no risk, no commitment.

Custom proposals are prepared after needs analysis. Flexible budget limits can be set per agent, user, and department.

Technical support, training, implementation consulting, and emergency support. A dedicated project manager is assigned to each customer.

Backed by 20+ years of enterprise experience. Step-by-step guidance throughout the onboarding process.

Minimum: 16 GB RAM, 8 CPU cores. Recommended: 32 GB+ RAM and a CUDA-capable NVIDIA GPU. On the system side, Windows Server 2019+ or Windows 10/11 with .NET Framework 4.8.

You do not have to inventory the hardware yourself: the platform scans the machine automatically (RAM, AVX2/AVX512, CUDA and VRAM, Vulkan) and flags models that will not fit. Details: Local AI Field Guide.

For some workloads yes, for others no — that is the honest answer. Document summarization, classification, tagging, embedding generation for semantic search, OCR and speech work fine locally. Long multi-step reasoning, complex code generation and very large context are better on cloud models.

That is why hybrid is the default: the same governance gate can route to a local model or a cloud model. What stays where is decided together with your own data during the POC.

Yes. The runtime binary, language model, embedding model, OCR, speech and image layers can each be fed from their own local source list; external internet access can be shut off entirely.

With localOnly on, cloud models are not listed and any cloud model request is refused categorically.

Yes — apart from hardware and electricity there is no per-use charge. The model runs on your own machine, so no request goes out.

This is not limited to chat: embedding generation, OCR, speech-to-text and text-to-speech workloads generate no token bill either.

No. The model runs in a separate child process; a model crash cannot take the platform down — your business modules keep running.

Crash-loop protection applies: after 3 crashes within 5 minutes the model is stopped, with guidance to try a smaller model or lower the GPU layer count. Idle models are unloaded from memory after 30 minutes by default.

Not while localOnly is on. This is a technical guarantee, not a policy statement: cloud models are not listed and cloud requests are refused.

Even in hybrid mode you can define a domain-based allow-list and a daily request quota. Cloud provider keys stay with the platform administrator, never descend into the customer workspace, and are always shown masked in the panel.