Frequently asked questions

Most of what you learn never gets synthesized — it sits in scattered PDFs, transcripts, notes, and threads you'll never re-read. stoa turns that pile into one connected, cited picture you can question, so your past reading compounds instead of evaporating.

Documents, today: PDFs, Word / PowerPoint / Excel files, plain text, Markdown, HTML, and CSV. And here's the point — you don't have to prep any of it. Drop it in however messy it is; stoa converts it to clean text, strips the noise (headers, boilerplate, reply-chains), and pulls out the meaning — the entities, the claims, how they connect — then builds its graph on that clean version. No formatting, tagging, or structure required from you. On the way: audio & video transcription, live-link fetching (paste a URL), richer image/diagram understanding, and source connectors. Want a specific format or connector? Tell us — we're actively building these and prioritize what people ask for.

No. Cleaning, extraction, graph-building, reconciliation, and retrieval all happen automatically. stoa shows you what it did if you're curious — but you never configure any of it. If you do want to peek under the hood, see our how it works page.

Two things most tools don't give you together. Your brain is a folder of plain files you own — the originals, the processed markdown, the wiki, and the graph — so you can move it, open it anywhere, or delete it anytime, with no lock-in. And you choose where the AI runs: a cloud model with your own key, or a fully local model, per brain.

Only if you choose it. A local brain runs entirely on your device. Point a brain at a cloud model and your prompts go to that provider with your key — but there's no silent cloud fallback: nothing leaves your box without a loud, explicit choice.

Always. Bring your own key (OpenAI, Anthropic, Bedrock, or anything LiteLLM supports), or run local models with Ollama. You're never locked to one provider — configure as many as you like and mix and match models across providers as you see fit. And if you'd rather not think about providers at all, the managed edition offers an option where we run the inference for you, fully private thanks to confidential computing — no one, not even us, can see your data.

RAG re-reads your raw files and re-derives the same context on every question — nothing accumulates. stoa compiles your sources into a persistent graph + wiki once, then maintains it. Retrieval fuses semantic and keyword search and walks the graph, so it finds connected facts a flat search misses.

Those are proprietary, closed-source systems: they often don't let you choose your own models or easily decide where your data is hosted, they're heavyweight, and they come with a hefty price tag. stoa is the opposite — completely open, with a transparent engine you can inspect, and you stay in full control of everything: both your data and the inference.

Every claim is cited to the exact passage it came from, and you can see the retrieval path that produced it — drilling down all the way to where each fact originated. When sources disagree or coverage is thin, stoa surfaces that instead of hiding it.

Yes. Invite people as viewers (read & ask) or editors (also add documents). It's private by default, and a shared link never grants access on its own — and collaborators can reach a shared brain over MCP too, from their own tools.

That's a core use case: a shared brain is one source of truth a whole team asks and contributes to — while you stay in control of who can read, who can add, and what agents may write.

It's for people who'd rather not run stoa themselves. We host it — synced across your devices, seamless team sharing, and always-on so your tools can reach it anytime — for a fraction of the cost of self-hosting, with zero ops. It's private by design: per-tenant encryption with your own key, never trained on, never retained, never shared, and production-grade reliability. See the Pricing page for the managed edition's full details.

Yes. By default you bring your own model provider, so anything you send goes only to the provider you chose. And with the optional Maximum privacy add-on, we run the inference for you inside a no-human-access architecture — so no one, not us, not even the model provider, can see your data. Rather than trusting administrators to follow privacy rules, the guarantee is enforced by the infrastructure itself, built on confidential computing:

No. You can keep hosting your brains exactly where they are. The managed edition is an option you can move to later — it never removes the self-host path or the files you own.