Frequently asked questions
What is brainattic?
brainattic is a knowledge base your team and your AI agents both read and write — under the rules you set. People work in a fast editor, agents work over MCP, and everyone retrieves the same governed source of truth.
How is it different from adding ChatGPT to our wiki?
A chatbot bolted onto a wiki can skim snippets but can’t safely write back, so your canonical docs drift out of date. brainattic is built the other way around: agents are first-class citizens that draft and revise the source itself through real tools — governed by human-curated rules, with transactional edits and a full audit trail.
Which AI assistants does it work with?
Claude, ChatGPT, Mistral, or any MCP client. You add brainattic as a custom connector with a single URL, and your whole team’s agents pick up the full toolset — each person scoped to exactly their own workspace.
If agents can write to our docs, what stops them from making a mess?
Three things: human-curated rules that every agent reads before it writes; transactional edits it can roll back; and an append-only audit log of every change. Governance is a first-class part of the product, not an afterthought.
How does search work?
Hybrid search fuses keyword and meaning with multilingual embeddings, down to the passage. Ask in plain language and get the right paragraph back — in any language — instead of a list of links.
Can we get our content out — and in what formats?
Yes, with no lock-in. Export a single document, a whole project, or your entire knowledge base to Markdown, HTML, or PDF. Markdown is your content verbatim, so what you put in is exactly what you can take anywhere.
Do you send our documents to OpenAI, Anthropic, or any AI model?
No. brainattic does not send your data to any third-party AI model. The AI you use connects to brainattic as your own client; the connector receives only the input needed for the specific action you ask for, and returns a scoped result.
Where is our data stored?
On infrastructure we control inside the European Union. Your data is processed and stored in Bulgaria (EU), with no third-party logging or analytics. Data only leaves the EEA through an optional integration you choose to enable.
How is our data kept separate from other customers’?
Per-tenant isolation is the foundation, not an afterthought. Every request is confined to your tenant, with logical isolation, least-privilege access, and per-user authentication, so nothing leaks across teams or companies.
Can we self-host, or is it only SaaS?
Most teams start on our managed, EU-hosted service. For organisations that need everything inside their own walls, brainattic can also be self-hosted on-premise as an enterprise option — including the search and retrieval pipeline, so nothing is sent to a third-party AI. Talk to us about an on-premise pilot.
How do we get started?
Pilots are open. Request one and we’ll set you up — connecting takes minutes: add the connector, everyone authenticates individually, and your agents get to work.
Still have a question? Email hello@brainattic.ai or request a pilot.