{
  "updated": "2026-05-31",
  "services": [
    {
      "name": "Managed Local AI",
      "url": "/managed-local-ai/",
      "summary": "Managed Ollama-based local AI on customer-owned or customer-rented GPU infrastructure with Open WebUI as the default interface.",
      "cta": "Order Managed Local AI",
      "href": "https://support.ezoshosting.com/cart/managed-local-ai/",
      "bullets": [
        "Ollama standard",
        "Open WebUI default",
        "RTX 4000 Ada class 20 GB model-fit shortlist",
        "Benchmark report before throughput promises",
        "No third-party AI API required by default",
        "Private RAG with source-answer smoke tests",
        "Local document and transcription workflow benchmarks",
        "OpenAI-compatible local API bridge benchmark",
        "Open WebUI RBAC and SSO/OIDC access scope"
      ],
      "plans": [
        {
          "name": "BYO Server Management",
          "from_usd_monthly": 299.18,
          "fit": "Customer-owned or customer-rented GPU server with managed Ollama/Open WebUI operations."
        },
        {
          "name": "Local AI Managed",
          "from_usd_monthly": 699.42,
          "fit": "Managed private local model hosting with benchmark-first setup."
        },
        {
          "name": "Team RAG",
          "from_usd_monthly": 999.6,
          "fit": "Document-assisted local AI and team knowledge workflows."
        },
        {
          "name": "Business Secure",
          "from_usd_monthly": 1499.9,
          "fit": "Controlled production rollout with security hardening, audit preparation, and support scope."
        }
      ],
      "runtime_policy": {
        "local_ai": "benchmark-first",
        "gpu_class": "RTX 4000 Ada class 20 GB systems are positioned for small-to-medium local models.",
        "readiness_gate": "Do not promise live local inference until GPU driver visibility, Ollama service health, and a model smoke test pass."
      },
      "current_model_shortlist": [
        {
          "use_case": "team_rag_embeddings_and_reranking",
          "candidates": [
            "Qwen3-Embedding 0.6B",
            "Qwen3-Embedding 4B",
            "Qwen3-Embedding 8B",
            "Qwen3-Reranker 0.6B",
            "Qwen3-Reranker 4B",
            "Qwen3-Reranker 8B"
          ],
          "positioning": "Multilingual retrieval, code search, source ranking, and private knowledge bases with measured storage, latency, and cited-answer behavior."
        },
        {
          "use_case": "document_intake_vision",
          "candidates": [
            "Qwen3-VL 8B",
            "Qwen2.5-VL 7B fallback",
            "Docling/OCR baseline"
          ],
          "positioning": "Private scanned PDF, screenshot, invoice, form, table, reading-order, and field-level extraction benchmark before workflow automation."
        },
        {
          "use_case": "code_assistant_benchmark",
          "candidates": [
            "Qwen3-Coder 30B",
            "Qwen3 14B fallback"
          ],
          "positioning": "Benchmark-only private repository assistant path on RTX 4000 Ada class 20 GB systems; context, concurrency, latency, and access controls decide fit."
        },
        {
          "use_case": "assistant_and_support",
          "candidates": [
            "Qwen3 8B/14B",
            "Gemma 3 12B"
          ],
          "positioning": "First-pass candidates for support assistants, summaries, drafting, and RAG answers before larger or specialized model trials."
        }
      ],
      "market_positioning": "Benchmark current Qwen3 Embedding/Reranker, Qwen3-VL, Qwen3-Coder, Gemma, Docling/OCR, OpenAI-compatible local API bridge, and Open WebUI access-control candidates on RTX 4000 Ada class 20 GB systems before promising throughput.",
      "offer_tracks": [
        {
          "name": "Private RAG with sources",
          "href": "https://support.ezoshosting.com/cart/managed-local-ai/Team-RAG/&step=0",
          "fit": "Internal documents, support archives, policies, and project knowledge with cited answers.",
          "stack_candidates": [
            "Open WebUI",
            "Qwen or Gemma chat candidate",
            "EmbeddingGemma or Qwen embedding trial"
          ],
          "smoke_test": "Representative corpus ingest, fixed benchmark questions, cited-answer requirement, VRAM, latency, and miss-behavior log."
        },
        {
          "name": "Local PDF and image extraction",
          "href": "https://support.ezoshosting.com/cart/managed-local-ai/Business-Secure/&step=0",
          "fit": "Invoices, forms, screenshots, and operational documents that need local extraction support.",
          "stack_candidates": [
            "Vision-language model benchmark",
            "OCR/PDF fallback",
            "Field-level validation notes"
          ],
          "smoke_test": "Real sample pages against expected fields, false-positive review, unsupported layout notes, throughput, and VRAM peak."
        },
        {
          "name": "Local transcription and meeting notes",
          "href": "https://support.ezoshosting.com/cart/managed-local-ai/Business-Secure/&step=0",
          "fit": "Interviews, internal meetings, and support recordings with private audio handling.",
          "stack_candidates": [
            "Whisper or faster-whisper benchmark",
            "English and German sample set",
            "Optional local summary pass"
          ],
          "smoke_test": "Representative 5 to 30 minute audio files, runtime factor, language quality, segmentation limits, and GPU use."
        },
        {
          "name": "Private AI app bridge",
          "href": "https://support.ezoshosting.com/cart/managed-local-ai/Business-Secure/&step=0",
          "fit": "Internal applications, scripts, agent tools, and prototypes that need a local OpenAI-compatible endpoint with controlled access.",
          "stack_candidates": [
            "Ollama OpenAI-compatible API",
            "Open WebUI RBAC/SSO scope",
            "Optional vLLM quantized serving trial"
          ],
          "smoke_test": "Run representative app calls, confirm access boundaries, log latency and VRAM, and document fallback behavior before production traffic."
        }
      ]
    },
    {
      "name": "AI Apps",
      "url": "/ai-apps/",
      "summary": "Private knowledge, team chat, workflow automation, and creative GPU app stacks managed around open-source tools.",
      "cta": "Compare AI Apps",
      "href": "/ai-apps/",
      "bullets": [
        "AnythingLLM or LibreChat",
        "Flowise and n8n options",
        "ComfyUI for creative GPU workflows",
        "vLLM only as optional advanced layer",
        "Team RAG benchmark path",
        "Qwen3-Embedding retrieval candidates",
        "Permissions and update-window scoping",
        "Code assistant benchmark path",
        "Qwen3-Coder 30B trial with fallbacks",
        "OpenAI-compatible local endpoint benchmark",
        "Open WebUI RBAC and SSO/OIDC scoping"
      ],
      "team_rag_fastpath": {
        "href": "https://support.ezoshosting.com/cart/managed-local-ai/Team-RAG/&step=0",
        "from_usd_monthly": 999.6,
        "benchmark_candidates": [
          "Qwen3-Embedding 0.6B",
          "Qwen3-Embedding 4B",
          "Qwen3-Embedding 8B"
        ],
        "positioning": "Private knowledge benchmark and deployment path; no live inference promise until GPU, Ollama, and target-model smoke tests pass."
      },
      "code_assistant_fastpath": {
        "href": "https://support.ezoshosting.com/cart/managed-local-ai/Business-Secure/&step=0",
        "benchmark_candidates": [
          "Qwen3-Coder 30B",
          "Qwen3 14B fallback"
        ],
        "positioning": "Private repository assistant benchmark; no live production claim until GPU, Ollama, and target-model smoke tests pass."
      }
    },
    {
      "name": "GPU Infrastructure",
      "url": "/gpu-infrastructure/",
      "summary": "Right-sized GPU servers for local inference, image workflows, and private automation stacks.",
      "cta": "Plan GPU Stack",
      "href": "/gpu-infrastructure/",
      "bullets": [
        "Dedicated setup",
        "Storage and backup planning",
        "Monitoring and maintenance",
        "RTX 4000 Ada class fit guidance",
        "Benchmark before performance promises"
      ]
    },
    {
      "name": "Open Source Hosting",
      "url": "/open-source-hosting/",
      "summary": "Managed open-source hosting with CyberPanel, domains, SSL, DNS, and human support.",
      "cta": "View Hosting Options",
      "href": "https://support.ezoshosting.com/cart/open-source-hosting/",
      "bullets": [
        "CyberPanel control panel",
        "WordPress, Nextcloud, Matomo and more",
        "No cPanel or CentOS claims",
        "Built for long-term maintenance",
        "Starter, Pro, Business, and Managed App paths",
        "Checkout remains final for term, options, renewal, taxes, and total"
      ],
      "plans": [
        {
          "name": "Open Source Starter",
          "from_usd_monthly": 12.91,
          "fit": "One small site, project, or association with maintainable open-source hosting."
        },
        {
          "name": "Open Source Pro",
          "from_usd_monthly": 29.91,
          "fit": "Active WordPress, Nextcloud, Matomo, or similar open-source applications."
        },
        {
          "name": "Open Source Business",
          "from_usd_monthly": 69.04,
          "fit": "Business sites needing predictable support, restore help, and operational review."
        },
        {
          "name": "Managed App Hosting",
          "from_usd_monthly": 99.06,
          "fit": "Managed application layer with updates, backups, monitoring setup, and maintenance planning."
        }
      ],
      "pricing_note": "Public starting prices observed in the HostBill cart on 2026-05-31; checkout remains authoritative for term, options, renewal, taxes, and final total."
    },
    {
      "name": "Domains",
      "url": "/domains/",
      "summary": "Domain registration, renewal, transfer guidance, DNS support, and secure live availability and price checks for project domains.",
      "cta": "Check live availability and price",
      "href": "https://support.ezoshosting.com/checkdomain/domains/",
      "bullets": [
        "Secure checker confirms final USD price",
        "Dated examples support quick comparison",
        "Transfers reviewed by registry rules",
        "DNS basics included"
      ],
      "pricing_note": "Popular TLD examples were observed in the public HostBill checker on 2026-05-31; the secure checker and checkout remain authoritative for availability, term, taxes, and final price."
    },
    {
      "name": "Support",
      "url": "/support/",
      "summary": "Human support, knowledgebase, client area, and ticket flow for managed hosting and AI stacks.",
      "cta": "Get Support",
      "href": "/support/",
      "bullets": [
        "Knowledgebase entry point",
        "Tickets for clients",
        "Managed setup consultation",
        "Real operational help"
      ]
    },
    {
      "name": "Affiliates",
      "url": "/affiliates/",
      "summary": "A partner program for people who recommend hosting and managed local AI they actually trust.",
      "cta": "Join Affiliates",
      "href": "https://support.ezoshosting.com/affiliates/",
      "bullets": [
        "Trackable partner flow",
        "No fake claims",
        "Eligible paid orders only",
        "Transparent program language"
      ]
    }
  ],
  "runtime_policy": {
    "local_ai": "benchmark-first",
    "gpu_class": "RTX 4000 Ada class 20 GB systems are positioned for small-to-medium local models.",
    "readiness_gate": "Do not promise live local inference until GPU driver visibility, Ollama service health, and a model smoke test pass."
  },
  "market_watch": {
    "last_checked": "2026-05-31",
    "sources": [
      "Ollama Qwen3 library",
      "Ollama Gemma 3 library",
      "Ollama Qwen3 Embedding library",
      "Open WebUI Ollama documentation",
      "NVIDIA RTX 4000 Ada specifications",
      "Ollama Qwen3-Coder library",
      "Ollama EmbeddingGemma library",
      "vLLM NVIDIA CUDA installation documentation",
      "Hugging Face Qwen3-8B",
      "Hugging Face Qwen3-Embedding-8B",
      "Hugging Face Qwen2.5-VL-7B-Instruct",
      "Hugging Face Whisper large-v3",
      "SYSTRAN faster-whisper",
      "Qwen3 Embedding official blog",
      "Hugging Face Qwen3-VL-8B-Instruct",
      "Docling official project"
    ],
    "operational_note": "GPU runtime is gated until driver visibility, Ollama health, and target-model smoke tests pass on the actual host.",
    "offer_track_note": "Next commercial tracks are benchmarked setup packages: source-backed RAG with reranking, local document/image extraction with Docling/OCR baselines, local transcription, and private code assistant trials. Claims remain gated by runtime and smoke-test evidence."
  }
}