Assistant
For your customers
An AI chatbot for your website, Slack, Teams and WhatsApp. It answers from your content with citations, books appointments, hands off to a human when needed and turns unanswered questions into leads.
Build an assistant →Ask your business anything.
An EU-hosted AI assistant that cites every answer. Type a question, or paste your website.
EU-hosted · every answer cited · free to start
scrollOne engine. Three products.
RAG Engine turns your website, documents and business data into AI answers your customers and teams can trust: every answer is grounded in your sources, cited inline, scored for confidence and written to an audit log. Hosted in the EU (Amsterdam). Your data is never used to train models.
Data Agents
For your team
Ask Google Analytics, Search Console, HubSpot, Salesforce, Zendesk, Snowflake, BigQuery or PostgreSQL in plain language and get answers backed by the actual numbers — with the query shown.
Explore data agents →Engine
For developers
The retrieval API underneath both: hybrid search (BM25 + vectors), neural reranking, a knowledge graph and agentic multi-step retrieval, with streaming, webhooks and SDKs. Bring your own OpenAI or Anthropic key.
See the engine →Answers you can audit
Every reply ships with a receipt, so a compliance officer, a lawyer or a support lead can check it in seconds. Drag the score to see what the assistant does at each level.
- Citations on every answer
Each statement links to the exact source passage — a page on your site, a PDF, a ticket or a row in your data. How retrieval works →
- Grounding score, 0–100
A confidence score on every answer. Low-scoring answers ask for an email instead of guessing. Self-improving answers → · Lead capture →
- Provenance log
Model, region, sources and timing are logged per answer, with PII redaction, audit logs and SSO on higher plans. Audit logs → · Trust centre →
RAG ENGINE · RECEIPT
- grounding
- 93 / 100 · high
- behaviour
- answered, cited
- citations
- 2 sources
- logged
- yes · eu-amsterdam
- next step
- none
keep for your records
VERIFIED
High. Answered with inline citations and written to the audit log.
Built for regulated and customer-facing teams
Retrieval models tuned per sector, with a live demo assistant on every industry page.
RAG Engine for
What does a first consultation cost?
$150 flat for 30 minutes — credited to your first invoice. 1
Book a consultation →
Everything you use flows into one engine
29+ integrations, from documents and chat to analytics, CRM and databases. Hover a node.
RAG Engine
Google Drive
Notion
Dropbox
OneDrive
Slack
Microsoft Teams
Discord
HubSpot
Salesforce
Pipedrive
Zendesk
Intercom
Freshdesk
Jira
Linear
Asana
GitHub
GitLab
Google Analytics
Search Console
Google Ads
Meta Ads
LinkedIn Ads
Snowflake
BigQuery
PostgreSQL
MySQL
WordPress & websitesHow it works
From a URL to a cited, logged answer in three steps — no engineering project.
- Connect
Paste a URL, upload documents or connect a source. Crawling, chunking, embedding and indexing run automatically — including JavaScript sites.
- Tune
Choose the model, retrieval settings and tone. Add verified answers, metadata filters and your own OpenAI or Anthropic key.
- Deploy
Embed the widget, connect Slack, Teams or WhatsApp, or call the API. Every answer is cited, scored and logged from day one.
Start free. No card.
€0Free — 1 assistant, 10 documents, 500 questions a month
€29Starter — 3 assistants, 50 documents
€49Pro — 10 assistants, 500 documents, API
€299Enterprise — SSO, audit logs, SLA, white label, on-prem
Bring your own LLM key on any plan. Prices per month; yearly saves about 17%.
See RAG Engine in action
Connect a source, ask a question, get a cited answer — in one short film.
People ask
Is my data used to train AI models?
No. Your documents power only your own assistants. RAG Engine never uses customer data to train models, and on any plan you can bring your own OpenAI or Anthropic key. Security →
Where is my data hosted?
In the EU, in Amsterdam. RAG Engine is built for GDPR from day one, with PII redaction, audit logs and SSO/2FA on higher plans. Trust centre →
How is RAG Engine different from Chatbase or CustomGPT?
Every answer carries citations, a grounding score and a provenance log you can audit; hosting is EU-based; and the same engine powers data agents and an API, not only a chat widget. RAG Engine vs Chatbase →
What can I connect?
Websites, PDFs and documents, Google Drive, Notion, Slack, HubSpot, Salesforce, Zendesk, Intercom, Google Analytics, Search Console, Snowflake, BigQuery, PostgreSQL and more — 29+ integrations. All integrations →
Does it work in my language?
Yes. Assistants answer in 50+ languages, and the interface is localized in English, German, French, Dutch, Portuguese and Spanish. Multi-language →
How much does it cost?
Start free with one assistant, 10 documents and 500 questions a month, no card required. Paid plans start at €29 per month. Pricing →