Blog Content → Brand Voice Extraction

Pull published WordPress posts, strip HTML, and create a Mavera brand voice profile capturing your blog’s tone and editorial style

Scenario

Your WordPress blog has years of published content that embodies your brand’s tone, vocabulary, and editorial style. You pull the most recent published posts, strip HTML to extract clean text samples, send them to Mavera Brand Voices to create a voice profile, then verify the voice with a test generation using the OpenAI SDK pattern.

Architecture

Code

Python

import os, re, requests
from openai import OpenAI

WP, MV = os.environ["WORDPRESS_URL"], os.environ["MAVERA_API_KEY"]
MB = "https://app.mavera.io/api/v1"
MH = {"Authorization": f"Bearer {MV}", "Content-Type": "application/json"}
strip_html = lambda h: re.sub(r"\s+", " ", re.sub(r"<[^>]+>", "", h or "")).strip()

resp = requests.get(f"{WP}/wp-json/wp/v2/posts", params={
    "status": "publish", "per_page": 50, "orderby": "date", "order": "desc",
    "_fields": "id,title,content,date,slug"
})
resp.raise_for_status()
posts = resp.json()
print(f"Fetched {len(posts)} published posts")

samples = []
for p in posts:
    body = strip_html(p["content"]["rendered"])
    if len(body) > 100:
        samples.append(f"{p['title']['rendered']}\n\n{body[:2000]}")
print(f"Prepared {len(samples)} samples")

voice_resp = requests.post(f"{MB}/brand-voices", json={
    "name": "WordPress Blog Voice", "samples": samples[:20],
}, headers=MH)
voice_resp.raise_for_status()
voice = voice_resp.json()
voice_id = voice["id"]
print(f"Brand voice created: {voice_id} — status: {voice.get('status', 'ready')}")

client = OpenAI(api_key=MV, base_url=MB)
test = client.responses.create(model="mavera-default", input=[
    {"role": "system", "content": "Write in the brand voice provided."},
    {"role": "user", "content": "Write a 2-sentence intro for a blog post about remote work productivity."},
], extra_body={"brand_voice_id": voice_id})
print(f"\nTest generation:\n{test.output[0].content[0].text}")

Example Output

{
  "id": "bv_4e8a1c3f",
  "name": "WordPress Blog Voice",
  "status": "ready",
  "sample_count": 18,
  "voice_attributes": { "tone": "conversational-authoritative", "vocabulary_level": "accessible-professional" },
  "created_at": "2026-03-17T10:15:22Z"
}

Test generation:
You don't need a corner office to do your best work — you need the right systems.
Here's how top remote teams are rethinking productivity without burning out.

Error Handling

REST API returns 404 or HTML instead of JSON

If /wp-json/wp/v2/posts returns a 404 or an HTML page, your permalink structure is set to “Plain.” Go to Settings → Permalinks and choose any other option (Post name is recommended). Flush rewrite rules by saving. Some security plugins (Wordfence, iThemes) block the REST API — whitelist your server’s IP or disable the REST API restriction.

Mavera 422 — insufficient samples

Brand voice creation requires at least 3 text samples with meaningful content. If your blog has few posts, supplement with page content (/wp-json/wp/v2/pages) or manually add representative copy. Each sample should be at least 100 characters of clean text.