Cloud server preconfigured with Ollama, Open WebUI and RAG. Run Llama 3, Mistral and Mixtral on your server — no per-message billing, no rate limit, no data sent outside.
Rollin Host AI Cloud Server is a VPS preconfigured with Ollama, Open WebUI, pgvector and Qdrant to run open-source LLMs (Llama 3, Mistral, Mixtral, Qwen) on your own server. 5 plans from US$ 12 to US$ 140/month, with up to 120 GB of RAM to run Mixtral 8x22B. Automatic 5-minute setup, OpenAI-compatible REST API, 100% private data (LGPD) and 24/7 human support.
Each plan comes with a VPS sized for the recommended model. You can switch models at any time.
no contract · cancel anytime
no contract · cancel anytime
no contract · cancel anytime
no contract · cancel anytime
no contract · cancel anytime
Self-hosted AI makes sense when privacy, predictable cost and independence matter more than GPT's latest feature.
Your data never leaves the server. Zero sending to OpenAI, Anthropic or Google. Critical for healthcare, legal, financial and any sensitive data.
You pay only the server. No per-token billing, no end-of-month surprises. US$ 12 or US$ 120 flat, regardless of 1k or 100M tokens.
Dedicated model, dedicated processing. Run big batches without waiting for OpenAI quota or paying premium tier.
Llama 3, Mistral, Mixtral, Qwen, DeepSeek — the whole open-source family runs natively via Ollama. Switch models in seconds.
Automatic setup: you sign up, get credentials, open Open WebUI and you're already chatting with AI.
Open-source model manager. Download, run and switch models with 1 command. Supports Llama 3, Mistral, Mixtral, Qwen, Phi, Gemma, DeepSeek and dozens more.
ChatGPT-like interface for your team in the browser. Saved chats, multi-user, document upload for RAG, shared prompts.
Vector database for RAG. Index your documents and the AI answers based on your content, citing the source. Plus+ has dedicated Qdrant.
Plus and Master include n8n integrated. Connect your AI to Gmail, WhatsApp, Sheets, CRM, ERP — visual workflows without code.
Ollama has 100+ models. The table shows the sweet spot per tier.
| Plan | Recommended models | Approx. speed | Use cases |
|---|---|---|---|
| Start | Llama 3.2 3B · Phi-3 Mini · Gemma 2B | 15-30 tok/s | Simple chatbot, data extraction, classification |
| Pro | Llama 3 8B · Mistral 7B · Qwen 2.5 7B | 8-15 tok/s | Customer support, RAG over docs, light agents |
| Plus | Llama 3 8B · Qwen 2.5 14B · CodeLlama 13B | 5-10 tok/s | Corporate RAG, multi-project, n8n automation |
| Master | Mixtral 8x7B · Llama 3 70B Q4 · Qwen 2.5 32B | 3-8 tok/s | Complex analysis, multiple models simultaneously |
| Enterprise | Llama 3 70B · Mixtral 8x22B · DeepSeek V3 | 2-5 tok/s | Compliance, isolation, audit |
Simple math: at how many tokens/month the server's fixed cost beats OpenAI's variable cost.
| Volume | OpenAI GPT-4o-mini | Rollin IA Cloud | Veredito |
|---|---|---|---|
| 100k tokens/month | US$ 1.5 | US$ 12 | OpenAI wins (low volume) |
| 1M tokens/month | US$ 16 | US$ 12 | Tied (Start covers it) |
| 10M tokens/month | US$ 160 | US$ 26 | Self-hosted Pro 6× cheaper |
| 100M tokens/month | US$ 1,600 | US$ 78 | Self-hosted Master 20× cheaper |
| Sensitive data | Not applicable | US$ 120 | Self-hosted Enterprise · only viable option |
| Feature | Rollin AI Cloud | Together.ai | Replicate | RunPod |
|---|---|---|---|---|
| Billing | Fixed monthly (US$ 12-140) | Per token | Per second | Per GPU hour |
| Data stays on | Your server (LGPD) | Their infra (US) | Their infra (US) | Allocated pod |
| Stack included | Ollama + WebUI + RAG | API only | API only | You install |
| Runs on CPU | Yes (all plans) | GPU only | GPU only | GPU only |
| BR billing | NF-e + PIX | USD | USD | USD |
| Human support | 24/7 | English only | English only | English only |
Another question? Open a ticket.
It is a VPS preconfigured with Ollama, Open WebUI, pgvector and Qdrant to run open-source LLMs (Llama 3, Mistral, Mixtral, Qwen, DeepSeek) on your own server. 5 plans from US$ 12 to US$ 140/month, with automatic 5-minute setup. You access via browser (ChatGPT-like) and via OpenAI-compatible REST API.
From US$ 12/month (Start, 12 GB RAM, Llama 3.2 3B) to US$ 140/month (Enterprise, 120 GB RAM, Mixtral 8x22B). Most popular is Pro at US$ 26/month — 24 GB RAM, runs Llama 3 8B with RAG via pgvector. No lock-in, no per-token billing.
Yes, slower than dedicated GPU. Llama 3 8B on CPU does 8-15 tokens/second — usable for chat and automation, slow for realtime streaming.
Worth it from ~1 million tokens/month (Start US$ 12 already matches US$ 16 of OpenAI at the same volume). At 10M tokens/month, Pro is 6x cheaper; at 100M tokens, Master is 20x cheaper. For sensitive data (LGPD), self-hosted is the only viable option.
Yes, at no cost. Ollama has 100+ models. Switch by command, only limit is disk space.
From the Pro plan it comes with pgvector. Upload documents, the system generates embeddings and indexes. AI answers citing the source.
Yes, from Plus it comes with n8n with native Ollama node. Common workflows: classify tickets, respond to leads, transcribe audios.
Full fine-tuning needs GPU. Light LoRA (3B-7B) runs on Master/Enterprise with a few training hours.
EVERYTHING: prompts, responses, RAG documents, history. Zero sending to third parties.
Yes. Ollama exposes a REST API 100% compatible with OpenAI. Just point the SDK to http://your-server/v1.
In 3 steps: 1) sign up for the plan (Pro is most common to start), 2) change the OpenAI SDK base_url to your server URL, 3) pick the equivalent Ollama model (Llama 3 8B replaces GPT-3.5; Mixtral replaces GPT-4 in many cases). Average migration time: 1-2 hours.
No. Ollama runs as an isolated process, monitored via systemd. Restarts on its own if it hangs.
Yes — Rollin Serviços Digitais e Tecnologia LTDA is a Brazilian company with a Tier III datacenter in São Paulo, NF-e, billing in BRL and 24/7 human support. First Brazilian cloud specialized in AI.
Comece em 5 minutos. Migração gratuita, suporte 24/7 em português e garantia de reembolso em 7 dias.
Usamos cookies para analisar o tráfego, melhorar sua experiência e personalizar conteúdo. Você decide o que aceitar — consulte a Política de Cookies.
Escolha quais categorias você permite. Os cookies necessários são essenciais para o site funcionar e não podem ser desativados.
Essenciais para navegação, segurança e funcionamento básico do site. Não rastreiam você.
Ajudam a entender, de forma anônima, como os visitantes usam o site (Google Analytics).
Permitem medir a eficácia de campanhas e exibir anúncios relevantes (Meta Pixel).
