Support • Product • Privacy
Frequently Asked Questions
Clear, plain-language answers about Chatwize: how it works, privacy & GDPR, setup, and integrations.
FAQ
12 questionsChatwize is an AI chatbot platform for websites that answers visitor questions using your approved sources (web pages, FAQs, and documents).
It helps reduce repetitive support work, improves response speed, and can hand off to a human when a case is complex or sensitive.
Chatwize is a cloud-based SaaS solution. Uploaded data and system data are stored on secure cloud infrastructure.
All storage stays within the EU. The database and document storage sit with Supabase in Frankfurt, the application runs on Fly.io in Amsterdam. The only processing outside the EU is OpenAI generating the answer, under EU standard contractual clauses. Backups run daily.
Chatwize runs on OpenAI models. Depending on your plan, you get GPT-4o mini, GPT-4o, GPT-5.4 nano, GPT-5.4 mini, GPT-5.4, GPT-5.5 and GPT-5.6 (Luna, Terra and Sol), with a maximum context of 16k.
We continuously improve our model options so we can select the best fit for quality, speed, and cost per use case.
The chatbot can answer questions covered by your approved sources (opening hours, services, policies, procedures, step-by-step guidance). If something is not covered, you can configure safe fallback messages or human handoff.
You provide knowledge sources (selected pages or documents). Chatwize uses those sources to produce answers. Update sources whenever information changes to keep answers aligned.
Yes. Chatwize answers in the language the visitor writes in, using modern OpenAI models.
Visitors can ask in their own language and receive an appropriate answer based on your approved sources.
Chatwize supports GDPR/AVG-friendly configurations such as a consent/disclaimer step and controls for what data is stored. Compliance also depends on your setup and policies, so align retention and legal texts with your requirements.
Yes. You can show a disclaimer (terms and privacy notice) that users must accept before chatting. This improves transparency and can support compliance requirements.
Yes. Chatwize can hand off when a visitor asks for a human, when confidence is low, or when the request matches rules you define.
Launch time depends on your content and configuration. Many teams go live after adding sources, testing common questions, and embedding the widget on the site.
Chatwize has two connection mechanisms. Webhooks forward conversations and leads to your own system. With function calling, the bot can call your API during a conversation, available on paid plans with a full model at 8k context or higher. There is no public Chatwize API for customers and no ready-made connectors: every route is set up manually.
Chatwize is not a good fit when there is no reliable source material, when every answer must be individualized professional advice, or when policy requires every message to be handled by staff.