Short answer: open Capacities, click into a daily note or object body, press ⌃+⌥+R (Mac) or Ctrl+Alt+R (Windows/Linux), speak the thought, press the hotkey again. AICHE inserts clean text right where the cursor is.
The Problem
Capacities is a thinking tool. The whole point is that you can pull up a Book, a Person, a Meeting, or a Project and dump context into it the moment it shows up in your head. That only works if capture is faster than thought.
In practice, capture is the slowest part. You open the daily note, the cursor lands, and now you're typing. By the time you finish the second sentence, the half-formed idea that made you open the app has flattened into whatever your fingers managed to commit. The block-based editor and the bi-directional links are great after the fact. They do nothing for you while you're still typing the raw thought.
What Changes
You stop typing the raw thought. You speak it.
Cursor in the daily note. Hotkey. Talk for thirty seconds about what's actually in your head. Hotkey again. AICHE transcribes, cleans up the filler, and inserts the result into the block you were already in. Capacities then does what it's good at: you mention a Person or a Project, you link it, and the backlinks panel takes it from there.
Speaking runs around 150 words per minute. Typing runs around 40. A two-paragraph daily note that takes four minutes to type takes about a minute to speak, and the version you spoke usually has more of the original context in it because you didn't edit it down mid-sentence.
How It Works In Capacities
- Open Capacities (desktop app or
app.capacities.ioin a browser). - Click into the daily note, an object body, or any property field that accepts text.
- Press ⌃+⌥+R (Mac) or Ctrl+Alt+R (Windows/Linux).
- Speak. AICHE is toggle-based, not push-to-talk, so your hands are free.
- Press the hotkey again to stop.
- AICHE inserts cleaned-up text at the cursor.
From there you do the Capacities part: type @ to link an object, # to tag, / for a block command. AICHE handles the prose. You handle the structure.
Daily Notes Without The Friction
Daily notes are where most Capacities users actually live. The promise is a low-pressure scratchpad you open every morning. The reality is that the friction of typing turns it into a checklist of three-word bullets.
With voice, the daily note becomes closer to a thinking log. You can describe what the meeting was about, what you decided, what's still open, why a particular decision felt off, all in the time it used to take to type a heading. Later, when Capacities surfaces Related Content against an object, there's actually something there to relate to.
A practical pattern: every morning, hotkey into the daily note, talk for sixty to ninety seconds about the day. Plans, what's blocking, what you're worried about. AICHE produces a paragraph. You add the @Project links by hand because that takes two seconds and it's the part that turns prose into a graph.
Object Bodies And Long-Form Content
The body of a Book, Meeting, or Project object is where Capacities expects depth. It's also where typing fatigue hits hardest, because depth means paragraphs.
Voice changes the economics. A Book object's notes section can hold the actual argument of the book in your own words, not a five-bullet summary you typed because typing more felt like work. A Meeting object can hold what was actually said and what you actually decided, not a sanitized headline.
AICHE does optional cleanup on the transcript: filler words removed, punctuation and paragraph breaks added. The output drops into the Capacities block editor as plain text, which means headings, toggles, and quotes still work the normal way - type / and pick the block - and the spoken paragraph sits inside whatever block you've chosen.
Properties And Short Fields
Most Capacities properties are short: a title, a one-line description, a tag value, a URL. Voice is overkill for a title. It's not overkill for the "Notes" property on a Person object, or a long-text custom property on a custom object type.
The workflow stays the same. Click into the long-text property, hotkey, speak, hotkey. AICHE inserts wherever the cursor was. Smart Insert puts text into whichever field is active, including custom properties on custom object types.
For non-English thinkers, AICHE's optional translation pipeline lets you think in your native language and have English text appear in the field. Useful if your second brain is in English but your first brain isn't.
Quick Capture Outside Capacities
Capacities has its own quick-capture integrations (Telegram, WhatsApp, email-to-inbox) for when you're away from the app. AICHE is a different layer: it works inside any text field on your desktop, including the Capacities web app, the Capacities desktop app, and any browser tab, chat, or editor where you'd otherwise paste later.
If you keep the Capacities web app pinned in a browser tab, the hotkey-and-speak pattern works the same way whether you're in the daily note, in an object body, or in the search bar looking for something.
Common Questions
Q: Does AICHE work in the Capacities web app, the desktop app, or both?
A: Both. AICHE inserts into whichever text field has the cursor on macOS, Windows, or Linux. The Capacities desktop app, the web app at app.capacities.io, and any browser tab all behave the same way.
Q: Will dictation interfere with @ object links and # tags?
A: No. AICHE inserts text only when you toggle the hotkey. The @ menu, # tag picker, and / block menu are Capacities' own UI and respond to keystrokes you type yourself. Speak the prose, type the link triggers.
Q: I take notes in German (or Spanish, or Japanese). Does AICHE handle that?
A: Yes. Speak in your native language. AICHE supports multilingual input and, with optional translation on, outputs English while you speak the source language. Useful if your Capacities space is in English but you think in something else.
Q: My notes have a lot of jargon, names, and acronyms. Will it mangle them?
A: Add them to Custom Vocabulary in AICHE settings. Repo names, internal project codenames, people, frameworks, anything you say often gets spelled the way you spell it.
Q: Is the hotkey push-to-talk or toggle?
A: Toggle. Press once to start, press again to stop. You don't have to hold anything down while you're thinking out loud, which matters for the longer object-body workflows where you might pause for a few seconds mid-thought.
Q: Does this replace Capacities' own AI Assistant?
A: No. They do different things. Capacities AI works on top of your existing notes - summaries, queries, lookups inside your space. AICHE is upstream of that: it gets the raw thought into the note in the first place. Use AICHE to capture, use Capacities AI to operate on what you captured.
Q: What about privacy? Capacities is GDPR-compliant and EU-hosted.
A: AICHE deletes audio within a few seconds of processing and keeps no audio logs. Local encrypted storage handles anything queued offline. See the privacy page and current feature docs for specifics.
Result: the daily note stops being a sparse list of three-word bullets and becomes the actual log of what you were thinking. Object bodies hold the depth Capacities was designed to hold. Capture stops being the slow step.
Try it now: open Capacities, click into today's daily note, press your hotkey, and talk for ninety seconds about what's actually on your mind right now. Look at what lands in the block.