> ## Documentation Index
> Fetch the complete documentation index at: https://docs.daoco.org/llms.txt
> Use this file to discover all available pages before exploring further.

# Learned brand voice

> How daoco learns the way your brand writes from your own published posts, and how to read, edit, and turn off that profile.

<Note>
  Learned brand voice is available automatically once the brand has enough
  imported post history.
</Note>

During onboarding you describe your voice in your own words: a few tone words and some writing notes. That is what you say you sound like. Once daoco has imported enough of your real posts, it also learns what you actually sound like, and writes from that.

## What it reads

Every day daoco imports the posts from your connected social accounts, together with the likes, comments, shares, and reach each one earned. See [Social platform import job](/core-workflows/social-platform-import-job).

Once a brand has at least **10** imported posts, daoco reads them and writes a voice profile covering:

* **Overall** — how the brand sounds, in a couple of sentences.
* **Sentence rhythm** — length, variation, fragments, line breaks.
* **Vocabulary** — the words and constructions you actually reach for.
* **Never says** — words, claims, and moves absent from every post.
* **Emoji** and **Hashtags** — your habits, including "none".
* **Recurring themes** — the subjects you keep coming back to.
* **Hook patterns** — how your posts open.
* **Per-platform differences** — only where a platform genuinely reads differently. Most brands have none.

It also keeps a handful of **exemplar posts**: real posts of yours that best show the voice, with the engagement each one earned. Drafting uses those as the rhythm to match, never as copy to reuse.

Only posts imported from your accounts are read. Posts daoco wrote for you are left out on purpose: learning your voice from daoco's own writing would drift the profile toward daoco rather than toward you.

Below 10 imported posts, no profile is written and your onboarding voice answers stay in charge.

## Where to find it

Open **Brand Hub** and select **Voice & Personality**. The learned voice appears first, followed by your initial voice setup.

## Editing it

Select a field to edit it, then choose **Save**.

An edited field is marked **Edited**, and daoco will not overwrite it. When the profile refreshes, your edited fields stay exactly as you wrote them and only the fields you never touched pick up the new reading.

**Removing a field.** Clear a field and save, and it is marked **Hidden by you**: it drops out of your voice profile and stays out. Refreshing will not bring it back.

**Handing a field back.** Edit the field and choose **Use learned version** to let daoco fill it in again.

Very long entries are refused rather than shortened. If a field is over its limit the save fails and the message names the field, so nothing you wrote is cut in half without telling you.

## Refreshing and turning it off

The profile refreshes about once a week, and sooner if a batch of new posts arrives. You do not need to do anything.

* **Refresh** rereads your posts now. Your edits are kept.
* The **Enabled** switch controls automatic updates for this brand. Turning it off keeps and continues to use the current profile. You can still choose **Refresh** when you want a new reading.

## What uses it

When the profile exists, it is used wherever daoco writes social copy for you: brand chat drafts, the post creation flow, and content automations. It outranks generic tone adjectives when the two disagree.

It is not used for research, analytics, image or video work, or SEO audits.

The [Daily Content Engine](/core-workflows/automations) automation is built on top of it.
