Your advanced AI interface, hosted in France
A powerful AI conversation frontend powered by SillyTavern, installed and maintained by DINAO. Characters, cards, prompts, and history stay on our French servers — sovereign when you connect a local model.
What is SillyTavern?
SillyTavern is a self-hosted chat interface dedicated to advanced dialogue with language models. Born in 2023 from a fork of TavernAI, it has become a major independent project, driven by hundreds of contributors. Its strength lies in the richness of its interface: character cards, lorebooks (WorldInfo), visual novel mode, fine prompt management, and a third-party extension ecosystem.
Above all, it is a frontend: it does not generate text itself but connects to an LLM backend. You can connect it to a local engine (KoboldAI/CPP, Ollama, or any OpenAI-compatible server) hosted on your DINAO instance, or to an external API (OpenAI, Claude, Mistral, OpenRouter…). It also integrates automatic translation, text-to-speech, and image generation via extensions.
Lightweight, it runs on any Node.js 20+ environment and deploys easily via Docker. DINAO preconfigures it, exposes it over HTTPS on your domain, and manages it for you, prioritizing a local backend to ensure data sovereignty.
Host SillyTavern at DINAO
Resource tiers compatible with SillyTavern prerequisites (minimum 1 vCPU / 512 Mo / 2 Go). Hosted in France, fully managed.
- 1 dedicated vCPU
- 2 Go RAM
- 20 GB NVMe
- Daily backups
- Managed & monitored by DINAO
- 2 dedicated vCPU
- 4 Go RAM
- 40 GB NVMe
- Daily backups
- Managed & monitored by DINAO
- 4 dedicated vCPU
- 8 Go RAM
- 80 GB NVMe
- Daily backups
- Managed & monitored by DINAO
- 8 dedicated vCPU
- 16 Go RAM
- 160 GB NVMe
- Daily backups
- Managed & monitored by DINAO
This application uses AI
The container hosts the application, not the AI engine (which requires dedicated GPUs) : inference is handled externally, with your own provider key. Prioritize a sovereign engine — Mistral AI, NumSpot, Scaleway, or OVHcloud AI Endpoints (France, GDPR) ; an international provider (OpenAI, Anthropic…) only if a specific capability requires it. Inference subscriptions are not included in hosting.
Technical details
You might be wondering…
What exactly is SillyTavern?
It is an advanced interface (frontend) for chatting with language models: character cards, role-play, lorebooks, fine prompt management, and extensions. It does not contain the model itself: it connects to an LLM backend, either local or via an external API.
Do my conversations stay in France?
Yes, if you connect SillyTavern to a model hosted locally on your DINAO instance: everything stays on our French servers. If you connect it to an external API (OpenAI, Claude…), the content goes to the provider, often outside the EU — see the sovereignty note.
Can I use my own models?
Yes. SillyTavern supports many local backends (KoboldAI/CPP, Ollama, OpenAI-compatible server) and external APIs. You choose your models and switch between them on the fly.
Do I need technical skills?
No for usage: the interface is ready to use, DINAO manages installation, the backend, security, and updates. A technical team can still adjust prompts, extensions, and the inference engine.
Can I change my plan or export my data?
Yes. You can upgrade or downgrade at any time, and your character cards, lorebooks, and history remain exportable — no proprietary lock-in.
