← Back to catalog Onyx · managed by DINAO

AI that knows your entire company

An AI search assistant powered by Onyx (ex-Danswer), connected to your docs, apps, and teams. Installed and maintained by DINAO — your knowledge stays on our French servers.

Hosted in FranceSovereign in local mode40+ connectorsGDPR CompliantOfficial publisher image
Overview

What is Onyx?

Onyx (formerly Danswer) is an open-source enterprise search and generative chat AI platform. It is the AI assistant connected to the documents, applications, and people in your organization : it indexes your internal sources and answers your questions in natural language, citing its sources.

Onyx maintains knowledge and synchronized access controls across more than 40 connectors (Google Drive, Slack, Confluence, Salesforce, GitHub, Notion…). It allows you to create custom AI agents with their own prompts, knowledge, and actions, and works with the LLM of your choice — local open-weight model or external API.

The Community Edition (MIT license) is freely available and can be deployed anywhere : workstation, on-premise, or cloud, via Docker or Kubernetes. The Enterprise Edition adds SSO (SAML and OIDC) and features useful for large organizations.

Compatible offers

Host Onyx at DINAO

Resource tiers compatible with Onyx prerequisites (minimum 4 vCPU / 8192 Mo / 50 Go). Hosted in France, fully managed.

Performance
4 vCPU · 8 Go · 80 Go
39,90 € /month excl. VAT
  • 4 dedicated vCPU
  • 8 Go RAM
  • 80 GB NVMe
  • Daily backups
  • Managed & monitored by DINAO
Order

2 hidden tier(s) (insufficient resources for this app) : Découverte, Standard

🧠

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.

Under the hood

Technical details

vCPU
4 vCPU
ideal : 8 vCPU
Memory
8192 Mo
ideal : 16384 Mo
Disk
50 Go
ideal : 100 Go
Image : docker.io/onyxdotapp/onyx-backend:latest Registry : docker.io Services : relational_db, index, cache, inference_model_server, indexing_model_server, api_server, background, web_server Ports : 3000/tcp
FAQ

You might be wondering…

Does Onyx make up its answers?

No. Onyx answers based on your indexed internal sources and systematically cites the documents used. You can verify each answer, which significantly reduces the risk of 'hallucination' compared to a generic chatbot.

Does Onyx respect the access rights of my documents?

Yes. Onyx synchronizes not only the content but also the permissions of each source. A user sees in the answers only the documents to which they already have access.

Which LLMs can I use?

Onyx works with any LLM. To remain sovereign, connect a local open-weight model hosted on your DINAO instance; you can also connect an external API with your own key.

What is the difference between Onyx and Danswer?

It is the same project: Onyx is the new name for Danswer. You benefit from the open source edition (MIT), with optional Enterprise features such as SAML/OIDC SSO.

Where are my data hosted?

On DINAO infrastructure in France, in one of the available datacenters. Your indexed documents and conversations do not leave the territory.