Short answer: open Tana, click into a node or daily page, press ⌃+⌥+R (Mac) or Ctrl+Alt+R (Windows/Linux), speak, press the hotkey again. AICHE inserts clean text right where the cursor is.
Why Tana Specifically
Tana is an outline where every node can carry typed fields. You create a #meeting supertag with an Attendees field, a Decisions field, and an Action Items field. Now every node tagged #meeting becomes a tiny database row inside an infinite outline. The outline is the interface. The supertags are the schema.
That schema part is fast. Adding a supertag is a # keystroke. Picking a field value is a click. But the prose that goes into those fields - the actual notes, the context, the paragraph that explains why a decision was made - that part is slow because it's typing.
A product review call ends. You create a node, tag it #meeting, and fill in the Attendees field (two seconds, it's a dropdown). Then you get to the Decisions field and stare at it. The decision was nuanced. It involved three trade-offs. Typing three sentences into a supertag field feels wrong, so you write five words and move on. The supertag structure is there. The content is not.
What Changes
You speak the content instead of typing it. Click into the Decisions field, press the hotkey, talk for twenty seconds about what was actually decided and why, press the hotkey again. AICHE transcribes, removes filler, adds punctuation, and inserts the text into the field. The supertag captures the structure. Voice captures the substance.
A real workflow: the product review call just ended. You open today's daily page in Tana.
- Create a new node.
- Tag it
#meeting. - In the Attendees field, type the names (fast, it's structured input).
- Click into the Decisions field. Press ⌃+⌥+R. Speak: "We're shipping the redesigned onboarding flow next sprint. Skipping the guided tour in favor of contextual tooltips because analytics show 70 percent of users dismiss the tour modal within three seconds. Chen is updating the tooltip copy by Thursday." Press the hotkey again.
- Click into Action Items. Hotkey. Speak: "Chen rewrites tooltip copy by Thursday. Sara updates the onboarding Figma with the removed tour screens. I'm filing the sprint ticket and tagging it with the Q3 activation OKR." Hotkey.
- The
#meetingnode now has structure and content. The live query you built last month - "all meetings where Action Items contain my name" - picks it up automatically.
Total time: about two minutes. Typing the same three fields would have taken six or seven, and the Decisions field would have been four words instead of three sentences.
How It Works in Tana
- Open Tana in your browser at
app.tana.inc. - Click into any node body, supertag field, or daily page.
- Press ⌃+⌥+R (Mac) or Ctrl+Alt+R (Windows/Linux).
- Speak. The hotkey is toggle-based, not push-to-talk. Your hands are free.
- Press the hotkey again to stop.
- AICHE inserts cleaned-up text at the cursor.
From there, you handle the Tana-native parts: # to apply a supertag, @ to mention another node, Cmd+K to run a command node. AICHE handles the prose. You handle the graph.
Daily Pages
Tana's daily page is the default landing surface. Most people treat it like a scratchpad with two-word bullets because typing paragraphs into an outliner feels heavy.
With voice, the daily page becomes a real log. Press the hotkey into a new node under today's date, talk for sixty seconds about what happened, what's blocking, and what's next. AICHE produces a paragraph. You add #standup if you want it queryable, or leave it as raw context. When you search for a project name next month, the daily page entry surfaces because the spoken version included the project name naturally - you didn't abbreviate it to save keystrokes.
Supertag Fields and Node Bodies
The body of a node tagged #project, #bug, #article, or any custom supertag is where Tana expects depth. It's also where depth rarely shows up, because typing three paragraphs into an outline node takes discipline.
Voice changes the math. A #bug node's Description field can hold the actual reproduction steps and the hypothesis about root cause, not a one-liner. A #project node's Context field can hold the full business case, not "see Slack thread." The supertag inheritance system means child supertags carry the same fields, so dictating into a #frontend-bug node (child of #bug) fills the same Description field with the same voice workflow.
AICHE's Content Organization handles paragraph breaks for longer dictation. A sixty-second stream about a project's status lands as two or three readable paragraphs, not a wall of text.
Live Queries and Search Surface
Tana's live queries filter nodes by supertag, field values, and date ranges. A query like "all #meeting nodes where Decisions field is not empty, from this week" only works if the Decisions field was filled in. Voice makes filling it in trivial, which means live queries actually return results instead of empty lists.
The same logic applies to Tana's search. Search is full-text across node bodies and field values. A five-word field is five searchable tokens. A dictated paragraph is fifty. The search surface grows proportionally to what you put in, and voice puts in more because it costs less time.
What AICHE Does Not Do
AICHE inserts text. It does not apply supertags, create fields, configure live queries, or trigger command nodes. You speak the content. Tana runs the structure. There are no voice commands, no wake words, and no Tana-specific grammar.
What You Get
- Unlimited voice notes with AI cleanup - filler words removed, punctuation and paragraph breaks added.
- Content Organization - long spoken streams arrive as readable paragraphs inside node bodies.
- Custom vocabulary - add your supertag names, project codenames, and team jargon. Spelled correctly every time.
- Multilingual voice input with auto-translation - think in your first language, get clean English text into the node.
- Zero-retention audio - audio discarded immediately after processing, within 1 second. No persistent copy.
Plans start at $3.99/mo (annual) with a 7-day free trial, no credit card. See pricing.
Common Questions
Q: Does AICHE work in Tana's web app?
A: Yes. Tana runs at app.tana.inc in a browser. AICHE inserts into whichever text field has the cursor at the OS level, so it works in Chrome, Firefox, Arc, Safari, or any browser on macOS, Windows, and Linux.
Q: Will dictation interfere with # supertag triggers or @ mentions?
A: No. AICHE inserts plain text only when you toggle the hotkey. The # supertag picker and @ node mention menu are Tana's own UI, responding to keystrokes you type yourself. Speak the prose, type the structure.
Q: Can I dictate into supertag field values like Date or Select?
A: Those fields are dropdowns and pickers, not free text. Dictate into text-type fields: Description, Notes, long text custom fields, and node bodies. Use Tana's pickers for structured field types.
Q: My supertag names and field names are unusual. Will it get them right?
A: Add them to Custom Vocabulary in AICHE settings. Internal codenames, project supertags, and domain-specific terms get spelled correctly every time after that.
Q: Is the hotkey push-to-talk or toggle?
A: Toggle. Press once to start, press again to stop. You don't hold anything down, which matters when you're talking for sixty seconds into a node body and want your hands free to scroll or glance at another tab.
Result: supertag fields stop being empty placeholders. Daily pages hold actual context instead of three-word bullets. Live queries return results because the nodes they filter on have content worth filtering.
Try it now: open Tana, click into a node body or supertag field on today's daily page, press your hotkey, and spend thirty seconds saying what the node is actually about. Read it back.