Adding knowledge

Everything your lens can answer starts with what you add. Mindola gives you one place to put it all: a single composer that accepts writing, voice, files, and links, and works out the format for you.

One composer for everything

On Home and on Knowledge, the composer is docked at the bottom of the page. Drop a PDF, audio, video, image, or Markdown file anywhere on the page, paste a link or a YouTube URL, or just write. Press Enter to save and Shift+Enter for a new line. The composer detects what you gave it, so you never pick a format first. Q&A pairs and bulk Markdown import live behind the + button on the left of the bar.

Speak instead of typing

Tap the microphone in the composer and talk. A living halo shows it is listening. Tap the halo to stop, or press Esc to cancel, and your words appear in the composer as editable text. Review, fix anything, and press Enter to save.

Voice notes are part of paid plans, with a daily transcription allowance that resets each morning. Your plan’s limits are shown in the app.

What each format becomes

However it arrives, Mindola handles each format the same way:

• Q&A (behind the + menu): a question paired with its answer. Great for facts you want phrased exactly your way.
• Written or pasted text: a thought, an article, or notes. Up to 2M characters per source; longer pastes are split into numbered parts automatically, so a 5M-character paste becomes three sources titled ”… (1/3)”, ”… (2/3)”, ”… (3/3)”.
• A link: a web page or a YouTube URL. Mindola pulls the content or transcript.
• PDF: read page by page; it becomes quotable with citations.
• Audio and video: transcribed, and the transcript indexed.
• Image: text in it is read with OCR so it is searchable.
• Markdown: a .md file or pasted Markdown. Headings and paragraphs are kept as chunks.

Import a whole folder of Markdown

The Markdown dialog has two modes. Paste is for one document. Import a folder is for many at once.

Switch to Import a folder and choose a folder on your computer. Mindola finds every .md and .markdown file inside it, including files in sub-folders, and each one becomes its own source. A title is taken from the file’s frontmatter, its first heading, or the filename.

A progress bar shows how far along the import is. Keep the window open until it finishes. You can press Stop import any time to halt it, and anything already imported stays.

Every file lands in the space you picked, so a folder import is the fast way to give a lens a large body of knowledge to answer from.

Where it lands

Short Q&A and brief notes become facts. Longer text, links, files, audio, and images become sources, chunked and indexed so the AI can search and quote them. See Facts and citations.

A short message confirms which one happened. Sources show “Indexing now…” for a few seconds while they embed.

If a source gets cut off mid-import (your laptop sleeps during a big upload, the browser tab closes, anything that interrupts the server), Mindola checks in the background. Sources that actually finished but never confirmed are marked Ready automatically. Anything that did not finish is marked Failed with a note to re-upload, so nothing sits in “Indexing now…” forever.

Choosing a space

Each dialog lets you pick a space, or leave the capture global. A capture with no space still feeds your private chat, but it will not reach a lens that is scoped to a specific space. See Spaces and Access and privacy.

Your Home screen

Home, the first item in the sidebar, is one calm column with three things:

• The capture composer, docked at the bottom. It is the same composer as on the Knowledge page.
• A pulse line under your greeting: visitors this week, questions asked, conversations all-time. Click it to open the full analytics. Before anyone has visited, it simply invites you to share your lens.
• Today, a short to-do list built from real signals. It never becomes a wall: a few items at most.

The Today list

Two kinds of tasks appear:

• Questions your lens couldn’t answer. When a visitor asks something your lens has to refuse, the question shows up here, up to three at a time. Press Answer, write the reply your lens should have given, and it is saved as Q&A knowledge on the spot. Next time, your lens answers it itself. Open conversation jumps to that visitor in the inbox if you would rather reply in person.
• Setup steps, like recording your voice, giving your lens a face or a greeting, or sharing your link. Each one takes you to the right place and disappears once it is done. Not now? Dismiss it with the ✕ and it stays away for a week.

The small ring in the corner tracks what you have handled today. Clear the list and it collapses to a single check.

Asking vs adding

• Ask when you want an answer. Open chat and type your question; the reply is grounded in your memory and the thread stays in your history. See Asking your brain.
• Add when you just want something in your knowledge for later. Use the composer on Home or Knowledge. No reply, no thread; it goes straight to your knowledge base.

Both end up in the same place: your knowledge.

Try this

• Drop tomorrow’s meeting notes into the Home composer on your way out. No filing, no format picking.
• When a refused question appears on Home, answer it there instead of opening the inbox. One reply teaches your lens forever.
• Big PDF? Split it into shorter files first. Smaller documents index faster and quote more precisely.