Writesonic Writing Style: Training a Voice from Files, Links and Text

Introduction

This guide walks through creating a Writesonic Writing Style (previously called brand voice) and training it from existing content, using the vendor’s own documentation at https://docs.writesonic.com/docs/brand-voice. It is documentation-based research, not a live API test. This article may contain affiliate links. The reader task is narrow: open the Writing Styles area, add training material by file, link or text snippet, then name the style so future generated content is tied to one brand identity. The documentation describes the mechanics; the ordering, checks and setup note below are our suggested process, not part of the vendor contract. One scope note matters from the start. The documentation states that access to this feature varies by plan (https://docs.writesonic.com/docs/brand-voice). Nothing here assumes a particular plan, price or quota. The home page describes Writesonic around an AI search growth engine with tracking, prioritization, action and measurement language (https://writesonic.com/), but that marketing framing is not a measured fact and does not change the Writing Style steps. Treat the Writing Styles page as a setup surface, not as proof that output will match your brand.

Documented Workflow

The documentation lays out the flow in two steps plus a naming step. Paraphrased, with each claim tied to its source:

1. In the left-hand menu, select Writing Styles, then click Get started. Existing Writing Styles appear on the Writing Styles page (https://docs.writesonic.com/docs/brand-voice).
2. Click new voice to create a new style, use the edit icon to change an existing one, or the trash icon to delete one (https://docs.writesonic.com/docs/brand-voice).
3. In the training step, import content by uploading a file, adding a link, or adding a text snippet. Accepted file types are PDF, DOC and DOCX, up to 10 MB per file. For links, click Add URLs and enter the webpage link, using the plus button to add more. For text, open the Add text snippet section and paste a snippet that reflects the brand’s distinctive tone (https://docs.writesonic.com/docs/brand-voice).
4. When finished adding files, links or text, click Next (https://docs.writesonic.com/docs/brand-voice).
5. Enter or edit the Style name. Each voice name represents a distinct brand identity or style, and the documentation says specific voice names help keep generated content aligned with the intended brand personality, tone and messaging (https://docs.writesonic.com/docs/brand-voice).

Our suggested process differs from the contract in one deliberate way: we treat the name as the last decision, not an afterthought. The name is the label that ties every later generation back to one identity, so it should describe the identity you just trained rather than the tool or the date.

Practical Walkthrough

Use this as an operational checklist. It reorders the documented blocks into one training pass and adds decisions the documentation does not make for you.

1. Decide which identity you are training before opening the page. One style should map to one brand identity or voice, because the documentation ties each name to a distinct identity (https://docs.writesonic.com/docs/brand-voice). If you write for two audiences with different tones, plan two styles rather than one blended style.
2. Open Writing Styles from the left menu and click Get started. Confirm you are looking at the Writing Styles page before creating anything (https://docs.writesonic.com/docs/brand-voice).
3. Choose new voice for a fresh style, or the edit icon if you are revising an existing one. Do not delete a style you still rely on; the trash icon removes it (https://docs.writesonic.com/docs/brand-voice).
4. Assemble sources before you upload. A workable mix is one authoritative document, one live page that shows the voice in context, and one short snippet of approved prose. The documentation supports all three import methods but does not say how they are weighted when they conflict (https://docs.writesonic.com/docs/brand-voice).
5. For the file route, upload a PDF, DOC or DOCX, up to 10 MB per file (https://docs.writesonic.com/docs/brand-voice). Prefer a document that is already approved and representative, not a draft that still contains rejected phrasing.
6. For the link route, click Add URLs, enter the webpage link, and use the plus button to add more links (https://docs.writesonic.com/docs/brand-voice). Choose pages that display the voice you want, since the page does not state how a link is weighed against a file.
7. For the text route, open Add text snippet and paste text that reflects the brand’s distinctive tone (https://docs.writesonic.com/docs/brand-voice). Keep the snippet short and clean; a pasted paragraph of approved copy is more useful than a long mixed document.
8. Click Next when the sources are in place (https://docs.writesonic.com/docs/brand-voice). This is the documented transition from import to naming.
9. Enter or edit the Style name (https://docs.writesonic.com/docs/brand-voice). Name it after the identity, not after the source file or the import date. If you trained a support voice and a launch voice, those are two names, not one.
10. Write a setup note for the style: the style name, and each source listed by type (file, link or text). This note is our reusable deliverable; the documentation does not require it, but it gives you a record of what the style was trained on.

Hypothetical worked example. Suppose an editor is setting up a style for a small software brand. Hypothetically, they upload a brand style guide PDF, add the company blog URL through Add URLs, and paste one approved paragraph into Add text snippet. They click Next, then name the style after the brand rather than after the guide file. The setup note reads: style name, one file (style guide), one link (blog), one text snippet (approved paragraph). This example is illustrative only; it is not a test result and does not predict output quality.

Reusable output. Keep the note in the same place as your content briefs. When someone asks why generated copy sounds a certain way, the note shows which identity was trained and from which source types. If the voice drifts, the note tells you what to revise or re-import, and it prevents two people from training the same identity twice under different names.

Completion Checks

Use observable checks, not impressions, before you rely on the style for generation.

1. The new style appears on the Writing Styles page under the name you entered (https://docs.writesonic.com/docs/brand-voice).
2. The training sources shown or recorded are the ones you intended: the file, link or text snippet you added, not a leftover from a previous attempt (https://docs.writesonic.com/docs/brand-voice).
3. The name matches one identity and is not a duplicate of another style (https://docs.writesonic.com/docs/brand-voice).
4. Your setup note lists the style name and each source by type, so the training pass is reproducible by someone else.
5. You have not assumed the style will fix a source problem. If the imported material was inconsistent, the page does not promise that training will resolve the conflict (https://docs.writesonic.com/docs/brand-voice).

These checks confirm setup, not output quality. The documentation does not describe a scoring or verification step for how closely generated text matches the imported sources, so any quality judgment is yours and should be made separately.

Limits

The documentation is specific about some things and silent about others. The silence matters for planning.

1. The page does not state how many sources a style can hold (https://docs.writesonic.com/docs/brand-voice). Do not plan around an assumed maximum.
2. It does not state how long training takes (https://docs.writesonic.com/docs/brand-voice). Do not promise a turnaround time to stakeholders.
3. It does not state how the tool weighs a file against a link, or a link against a text snippet, when they conflict (https://docs.writesonic.com/docs/brand-voice). Keep sources consistent so you are not relying on an undocumented tie-break.
4. Access varies by plan, per the documentation note (https://docs.writesonic.com/docs/brand-voice). Confirm your own access rather than assuming the feature is available.
5. The home page markets an AI search growth engine with tracking and measurement language (https://writesonic.com/). Those are marketing claims, not measured facts, and they are separate from the Writing Style setup described in the documentation.
6. This guide is documentation-based research, not a live API test. It does not verify behavior beyond what the cited pages state.
7. The absence of a detection or verification signal on the page is not proof of anything about authorship or output. Do not treat it as evidence that generated content is human-written or brand-accurate.

Final Verdict

Writing Style setup is a short, bounded task: open Writing Styles, start a new voice or edit an existing one, import a file, link or text snippet, click Next, and name the style after one identity. The documentation supports those mechanics directly (https://docs.writesonic.com/docs/brand-voice). What it does not support is any promise about weighting, training time, source limits or output fidelity. Treat the setup as the controllable part and the output as something you still review.

The practical value is the ordering: decide the identity first, assemble consistent sources second, name the style last, and keep a setup note that lists the style name and each source by type. That note is what makes the style auditable later, when someone asks which identity was trained and from what. Keep the plan modest, keep sources consistent, and verify access under your own plan before depending on the feature.

Official Sources

• Writesonic documentation, Writing Style (previously brand voice): https://docs.writesonic.com/docs/brand-voice
• Writesonic home page: https://writesonic.com/