Brain

Brain is your personal knowledge base for the Volnyn assistant. Add your documents once, and chat, agents, and workflows can search what you know — not just the open web.

Your assistant can already build websites, draft content, and run automations. What it cannot do out of the box is know what your business knows: your policies, price lists, meeting notes, product docs, and where any of that lives.

Brain is where that knowledge lives.

You add files, paste text, or fetch a URL. Volnyn extracts the text, chunks it, and embeds it. The assistant can then search across all of it and answer from your real content — with citations — instead of guessing.

Open Brain from the left sidebar at Dashboard → Brain (/user/brain).

How Brain fits with chat and skills

Brain is the knowledge piece of what makes the assistant useful. Each answers a different question:

They work together. An agent might search Brain for your pricing policy, follow a skill to format a quote, then help you turn that into a page or email.

The mental model for the rest of this page: indexing a document makes knowledge available; once it is Ready, chat and agents can use it automatically.

Where you access Brain

Open Brain from the left sidebar in your Volnyn dashboard, or go to /user/brain.

On the Brain page you will see:

• A short stat line — how many documents you have, how many are indexed, how many passages (chunks) exist, and any failures.
• An Add box with three tabs: File, Paste text, and From URL.
• Your document list with status, size, and preview.
• A Try it search box so you can preview what Brain returns before trusting an answer in chat.

If document indexing is not configured on the server (no embedding key), the page shows a warning and uploads cannot be indexed until that is fixed.

Adding a document

A document is anything you want the assistant to be able to search later. Volnyn reads the text, indexes it, and keeps it available to chat, agents, and automations.

File upload

1. On the Brain page, open the File tab.
2. Choose a file (PDF, Word, text, CSV, Markdown, JSON, or an image of a document).
3. Click Add.

Limits:

• Up to about 20 MB per file.
• Supported types include .pdf, .txt, .md, .csv, .json, .log, .doc, .docx, and common image formats (.png, .jpg, .jpeg, .webp) when text can be extracted.

Paste text

1. Open the Paste text tab.
2. Paste a policy, notes, a price list, or any other prose.
3. Optionally add a title.
4. Click Add.

If you skip the title, Volnyn uses the first line of the pasted text (trimmed) as the document name.

From URL

1. Open the From URL tab.
2. Paste an https://… link.
3. Click Fetch.

Volnyn downloads the page, strips scripts and styles for HTML, extracts readable text, and indexes it. Give it a clear title if the URL alone would be hard to recognize later.

What gets indexed

Brain indexes the text in your content:

Empty extractions are rejected — if Volnyn cannot find readable text, the upload fails with a clear error instead of creating a blank document.

What happens after you add a document

Once a document is added, Volnyn takes over. You do not manage chunking or embeddings yourself:

1. Extract — text is pulled out immediately so you know right away if a file is unreadable.
2. Queue — the document is created with status pending and a background job chunks and embeds it.
3. Ready — when indexing finishes, status becomes ready and the document is searchable in chat and tools.
4. Failed — if embedding fails, status becomes failed with a short error. You can reindex after fixing the cause.

How long until it is searchable? Usually from a few seconds to a couple of minutes, depending on document size and server load. The status dot on each row shows progress; it turns green when the document is Ready.

Managing your documents

The document list

Each document shows:

Document actions

On each row you can:

Statuses

Team sharing

Documents can be shared with your team when Team Scope is enabled for your account.

• Documents you own stay under your control for edits and deletes.
• Teammates may see and search shared documents according to team visibility rules.
• Only the owner or a team admin can reindex or delete a teammate's document. If you try, Volnyn returns a clear permission error.

Your indexed content is used only to answer your (and your team's) assistant — not to train third-party models as a product feature of Brain.

Giving the assistant knowledge

Adding a document to Brain makes it available. You do not need a separate attach step for every chat: once a document is ready, the assistant can call the built-in search_knowledge tool whenever a question looks like it is about your content.

That same tool is available from:

• Chat — ordinary assistant conversations.
• Agents / Claw — agent runs that need company-specific facts.
• Automations and workflows — nodes that look up your docs mid-flow.

The tool returns matching passages with the document title and a relevance score. The assistant should quote and cite those passages rather than inventing from general knowledge. If nothing is indexed yet, it should tell you to add documents on the Brain page. If nothing matched, it should say so instead of guessing.

Prompting the assistant to use Brain

Once documents are Ready, use the assistant normally. It reaches for Brain when a question is about your internal content.

Prompts that work well:

• "According to our internal docs, what is our refund window?"
• "Find the launch retro notes and summarize the top three action items."
• "What does our pricing page say about annual plans?"
• "Search our knowledge base for the onboarding checklist and turn it into an email."

Nudge the assistant toward Brain when you want a grounded answer: phrases like "according to our docs," "search our knowledge base," or "what do we know about" make it clear you want a cited answer from your content, not a general one.

Try search on the Brain page

Brain search is not limited to chat. On /user/brain, use the Try it box to run the same hybrid search the assistant uses.

• Enter a question (full questions work better than bare keywords).
• Review ranked passages with title, snippet, and score.
• Use this to verify that a policy or price list is indexed the way you expect before relying on it in a customer-facing reply.

FAQ

What is the difference between Brain and Skills?

Brain is knowledge your assistant can search (documents, notes, files). Agent Skills are instructions that teach an agent how to do a task your way. Use Brain for "what do we know about X," and skills for "here's our process for doing X."

Do I need to code to use Brain?

No. Upload a file, paste text, or fetch a URL. Volnyn handles extraction, chunking, and embedding.

Can the assistant take actions in my files?

No. Brain is read-only for your knowledge. It indexes content so the assistant can search it. It does not edit or delete files on your computer or in third-party apps.

How is this different from web search?

Web search finds public pages. Brain finds your uploaded and fetched content — policies, notes, and docs the open web cannot see. Prefer Brain for anything about "our," "my," or "the company's" specifics.

How long until a document is searchable?

It depends on size — often from under a minute for a short note to longer for large PDFs. The row shows indexing… until status flips to ready.

What happens if indexing fails?

The document shows failed with a short error. Fix the cause (for example, wait until embeddings are configured, or re-upload a readable file), then use Reindex.

Can I remove a document?

Yes. Delete it from the Brain list. All indexed passages for that document are removed. This cannot be undone.

Is my content used to train models?

Indexed content is used to answer your assistant and team searches. Brain is not a product feature that ships your documents into public model training.

Who can see my documents on a team?

It depends on team sharing. Shared documents are visible to teammates under Team Scope rules. Only the owner or a team admin can change or delete someone else's document.

What file types are supported?

Common document formats such as PDFs, Word (.doc / .docx), plain text, Markdown, CSV, JSON, logs, and many images of documents, up to about 20 MB per file. If Volnyn cannot extract readable text, the upload is rejected.

Can I preview search results without chatting?

Yes. Use Try it on the Brain page to run the same search the assistant uses and inspect ranked passages.

Next steps

• Open Brain in your dashboard and add your first policy or product doc.
• Ask chat a question that only your docs can answer, and check that it cites Brain.
• Package a recurring process as an Agent Skill on top of what Brain already knows.