🗣️ How voice memos and auto-fill works with the new Tana Mobile iOS app

The default destination from the Capture tab is now Today (but possible to change destination when capturing).

You can now navigate to any node in the Browse tab and capture into it.

You can now select a Supertag when you record the voice memo.

If you apply a Supertag, the fields will be shown in the voice recorder (so you know what to cover).

Auto-fill will attempt to fill in the fields based on the recording. The whole transcript will be shown below the fields.

The output will be a lightly formatted transcript (removing filler words).

For longer transcripts, AI will generate a title based on the content – If the voice memo is very short the title will be the transcript.

The audio file and raw transcript will always be available in source material that can be easily accessed from node options and a new audio icon next to the node.

The “Rewrite” feature you may have used on voice memos, will now create a temporary draft, that you can choose to move into the node or your notes if you are happy with it.

The intention is to allow you to experiment with Rewriting the content, but only keep what you are happy with.

The new app has a temporary limit of 1 hr for voice recordings (we’re working to increase this!).

You will see a notification when closing in on the limit.

A) I want to keep using the Inbox as destination, but with the new processing

B) I want to keep using the Inbox with my own command/processing

If you want fields to be auto-filled, you should say the name of the field first, e.g. “Working on today – I need to buy a new TV for the office, and review feedback on the new app. General mood – Excited about the launch!”

The processing may try to fill in fields you haven’t addressed explicitly based on the content in the recording, if you would like to avoid this, see below.

If you want to specify how AI should treat a field, you can try adding in prompts in the AI instructions of the field in Field config:

🚨 Known issue: Field auto-intialization is not working yet, so no fields in nodes created from mobile will be auto-intialized – but they may be auto-filled if you cover them in the recording.

Our intention with this change has been to shorten the path from voice to value – allowing you to focus on the output, not the audio recording in itself. But we always want you to be able to go back to listen to the original recording if you need to, download it, play around with extracting things from it.

The source material is easily accessible with a new voice icon button on desktop and mobile, both from collapsed nodes and from the node options menu.

The Rewrite as function has existed for some time, but has gotten a slight revamp. Rewritten content will now be shown as a draft in a new panel, and will disappear unless you choose to keep it.

To keep it, you can drag it into the output node, move it to a location or add it to Today.

It is not possible to publish drafts, you will need to store it first.

When we asked our users, around half of them said they wanted to have captured content sent to Today, and many have set up workflows to move it over automatically.

Having content sent to an Inbox requires the user to manually process these, tag them and move them to the right location. With the possibility to choose destination and tag in the mobile app, we think more of this can happen in the moment of capture, rather than after, reducing the need for an Inbox.

Once the app is approved in App Store, all voice memos will be processed with AI, which requires a Tana Core subscription.

During the Testflight community beta period, all users will be able to test voice memos, without having a Tana Core subscription.

15 min of voice memo transcription will require ca. 50 AI credits (5000 credits included per month in Tana Core subscription = 25 hours of voice memo transcription). In addition auto-filling will require some credits, how much depends on the content.

Using AI chat, the “Rewrite”, or other AI commands, on the source material after transcription, will require additional credits.

No, this is no longer possible. Our focus is on turning voice into structured content in Tana, we will not be an ideal service for pure audio recordings.

Not at the moment. If this is something that is important, we’d love to hear about it.

Yes, if you want to have content sent to the Inbox, this will still be possible.

1. In the current version, you need to pin the Inbox to make it appear in destination picker (Zoom into the Inbox, drag it to pinned).

2. You will need to select the Inbox as destination when capturing things in the mobile app.

We know this is not ideal, and will be looking into how we can make this easier for those who always want to capture to the Inbox.

Yes, all voice memos will be transcribed with the new processing, regardless of which node they’re sent to.

The exceptions are the Inbox or nodes that have custom Audio commands set up. Scroll down to the section on technical details at the bottom.

Yes, existing commands will continue to work as before, but be aware that these will require activity on the desktop client to run the processing, so the output will not be instantly available on mobile.

The old voice memo processing on Inbox was attempting to detect multiple items in one recording and would show these as tagged items in a tab. This is not yet supported in the new processing, but if you leave the Inbox unchanged, you will still get this behavior if you capture voice memos to the Inbox (see description above for how to capture to the Inbox).

If you’re using Ctrl/Cmd+K > Capture Voice on desktop, you can do Ctrl/Cmd+K > Set transcription language (recommend selecting auto-detect if switching between languages)

🚨 The difference between keeping your old processing and the new, is that old processing will require activity on the desktop client to run, while the new processing will make transcript instantly available in the mobile app after recording.