The RAG system for your data: sources in, answers with sources out.
RAGSuite is the knowledge engine for your data. Connectors read in your documents and apps, the engine retrieves and ranks the matching passages and answers with a source: to your team in chat and search, and to AI tools through MCP and API. You choose the language model.
Rather self-host? Install the Community Edition →
AI Search: results with an AI overview and source, as a search widget on your website or intranet.AI Assistant: chat with your company knowledge, every answer with a source, in the widget and the app.AI Connectors: plug in your apps, and use your knowledge in Claude Desktop or Cursor via MCP.
What is an enterprise RAG platform?
RAG stands for retrieval augmented generation. Put simply: before the AI answers, the engine looks up the matching passages in your own documents and writes the answer from them, with a pointer to the source. So the AI answers from your company knowledge, not from generic internet knowledge.
Whether you call it RAG software, a RAG solution or a RAG platform for companies, this is the principle. Building one with RAGSuite means: create an account or install, connect your sources, pick a model: no development needed.
Retrieve
The question is matched against the index of your connected sources.
… 28 days of leave in year one …
… request two weeks ahead …
Rank
The matching passages are found and ranked by relevance.
Answer
The language model writes the answer and names its sources. It goes out to people and to AI tools.
Learn more: What is RAG? For business leaders →
The engine also hands its knowledge to AI tools that speak MCP. MCP server → · What is a knowledge engine? →
From your sources to answers with sources.
Bring in your content.
Websites, PDFs, Word documents: what your team already uses.
Crawl (depth 1–5) + upload with preview · PDF, DOCX, TXT
Connect your sources.
Pull knowledge from the apps where it’s created: over an open standard.
SharePoint, Slack, Confluence, Google Drive, Notion, Gmail, n8n · MCP server & client
Answer where your team works.
AI Search and AI Assistant on your website, intranet or app, every answer with its source.
Streaming widgets with voice input and read-aloud · REST API · mobile app for iOS & Android
Get better.
See what people ask and how good the answers are.
Feedback, history and dashboard · Enterprise adds model comparison, query tracing and full audit logs
Connectors, engine, AI search and AI chatbot, in one platform.
Connectors bring sources into the engine; AI search and the AI chatbot give answers to your team. Everything is in every plan, including the free Community Edition.
AI search
The search front end of the engine: search across every document and app, as a widget for your website or intranet. Every result links to its source.
The notice period is three months to the end of a quarter. 1 Price increases above 5% trigger a special termination right. 2
… may be terminated with three months’ notice to the end of a quarter …
AI chatbot
The chat front end of the engine: a chatbot that answers from your content, with streaming, voice input, read-aloud and a source on every answer. You set up your own assistants for teams and tasks without code.
Connectors
One way in for every source. SharePoint, Slack, Confluence, Google Drive, Notion, Gmail and n8n are connected; with MCP you add further tools, and AI tools that speak MCP can query the engine.
In detail:AI searchAI chatbotBuild AI assistantsConnectorsSharePointConfluenceWebsite chatbotAI knowledge baseMCP serverKnowledge engine
Not a sales pitch. An honest assessment.
- You tell us what you need20 minutes, online
- We show it on a setup like yoursYour sources, your deployment
- You get a clear recommendationCommunity is enough, or Enterprise pays off
What’s in Community, RAGSuite Engine and Enterprise?
The Community Edition is a complete product under Apache 2.0. RAGSuite Engine and Enterprise add what a company needs to govern it: RAGSuite Engine run by us in Germany, Enterprise on your side.
| Feature | Community | RAGSuite Engine | Enterprise |
|---|
Pricing and plans in detail: See pricing →
Answers that get better over time.
Every answer has a source, your team gives feedback, and the dashboard shows what people ask. In the RAGSuite Engine and Enterprise you also compare models side by side and trace every query in detail.
Three months to the end of a quarter. Price increases above 5% allow special termination.
The notice period is three months to the end of a quarter.
Roles, sign-in and evidence for the whole organisation.
With RAGSuite Engine and Enterprise you map organisations, teams and roles, sign users in with Google, keep the full audit log with export, and set retention rules and compliance exports.
Every major language model, or a local one.
RAGSuite isn’t tied to one model provider. In the RAGSuite Engine you choose the model per project; the plans include credits for model usage, or you use your own API key (BYOK). Self-hosted, you use a model with your own key, an EU-based provider or a local model via Ollama.
Build a RAG system in minutes, not months: the way that fits you.
Hosted in Germany
Create an account, connect sources, done. We run RAGSuite in a data centre in Germany.
RAGSuite Engine →Self-hosted
Install with a public CLI, native or Docker, on your servers or with a host of your choice.
Self-hosted AI →Air-gapped
With a local language model and no internet connection at all, licence activated offline.
On-premise & air-gapped →Self-hosted in three commands
$ npm install -g @ragsuite/ragsuite@latest $ ragsuite init $ ragsuite start ✔ api :9090 · web :9191 running
“When is the next staff meeting?”
On 14 November at 10 am in the canteen.
Intranet · EventsAnswers with sources, in the app and by voice.
The mobile app for iOS and Android signs in to your RAGSuite, in the RAGSuite Engine or on your server, and comes with every plan. In the widgets, users ask by voice and have answers read aloud.
Questions about the enterprise RAG platform.
What is an enterprise RAG platform?
An enterprise RAG platform connects a language model to your company knowledge. Before each answer it looks up the matching passages in your documents and apps and answers from them, with the source. RAGSuite combines AI search, an AI chatbot and connectors in one platform, in the cloud or on your server.
Which sources can RAGSuite connect?
Websites via crawling, uploaded documents such as PDF, DOCX and TXT, plus SharePoint, Slack, Confluence, Google Drive, Notion, Gmail and n8n. With MCP you connect further tools. Every connector is in every plan, including the free Community Edition. All content goes into your own search index.
Which language models work with RAGSuite?
All the common ones: large models from leading providers, EU-based providers and, self-hosted, local models via Ollama or any OpenAI-compatible endpoint. You choose the model per project and can switch at any time. That lets you balance cost, quality and data protection per project.
What's the difference between Community, Cloud and Enterprise?
Community is the complete platform, free under Apache 2.0, that you run yourself. RAGSuite Engine adds the full feature set, such as roles, Google sign-in and full audit logs, run by us in Germany, in plans from €39 a month. Enterprise brings the same features to your servers, price on request.
Do we need our own developers to run it?
No. In the Cloud you need no IT of your own. Self-hosted, one IT person can install it: the CLI sets RAGSuite up natively or with Docker, and ragsuite doctor checks the prerequisites. Configuration happens in the dashboard, without code. For deeper integration there are the REST API, webhooks and MCP.
What does retrieval augmented generation mean?
Retrieval augmented generation means the AI first fetches matching information (retrieval) and then uses it to write its answer (generation). So the answer rests on your documents rather than only on the model's training data. Because the source comes with it, your team can check every answer.
Try the platform with your own documents.
7 days of RAGSuite Engine free, no credit card, or we show you the platform in a demo on a setup like yours.