What 99 Languages Means for Your Content Reach
Your Google Search Console data can show real search demand in a language your site has never published a word of. Almost nobody checks for it, and the ones who do rarely have a fast way to act on what they find.
Before going further: BrassTranscripts, the transcription tool referenced throughout this post, is built by Copper Sun Content and Creative, the same company behind Brass-SEO. Two products, one company, same instinct — take a task that normally needs an expert and hand back a plain answer. Worth knowing that up front, before you click through to a product you've never heard of and wonder why we're pointing you there.
Quick Navigation
- The Demand You Are Not Checking For
- What 99 Languages Actually Means
- Why This Is a Content Problem, Not an Audio Problem
- Finding the Signal in Your Own GSC Data
- From Signal to Published Content
- See It Happen for One Real Business
- What This Does Not Solve
- Frequently Asked Questions
The Demand You Are Not Checking For
Google Search Console records impressions by country for every query your pages appear in, including countries where the dominant language is nothing like your site's. Brass-SEO can pull that country-by-query breakdown directly, which is where a language gap first becomes visible.
Here's the part most site owners skip: a page can rank and generate impressions in a country before anyone has written a word aimed at that audience. Shared technical terms, a brand name that reads the same across languages, or a partial match on a query can put your page in front of searchers you never targeted. GSC logs the impression either way. Most business owners never filter by country, so the impression sits in the data, unseen, month after month.
This isn't a hypothetical. It requires two separate capabilities working together: the data to spot the signal, and a way to produce content in that language once you find it. Most businesses have neither set up to talk to each other. GSC data lives in one tab. Content production lives somewhere else entirely, usually with no language capacity beyond whatever the founder speaks.
What 99 Languages Actually Means
BrassTranscripts advertises 99+ language support, and its own guide to non-English transcription is upfront that accuracy is not the same across all 99 — it's tiered, and the tier tracks how much training data each language had going into the underlying model.
The guide sorts languages into three accuracy bands:
| Tier | Languages | What the Guide Says |
|---|---|---|
| Tier 1 — professional-grade | English, Spanish, German, French, Italian, Portuguese, Japanese | Abundant training data. English alone handles American, British, Australian, Indian, South African, Irish, and Scottish dialects well. |
| Tier 2 — good accuracy | Dutch, Polish, Russian, Korean, Mandarin, Arabic, Hindi, Turkish | A slight step down from Tier 1. Brazilian Portuguese gets a specific callout for extensive training data. |
| Tier 3 — variable | 80+ remaining languages | Limited training data. The guide's own words: "expect some words or phrases to be mistranscribed." Human review recommended before publishing. |
The guide argues Tier 3 still earns its keep: even 70% accuracy "eliminates most manual work," in its own words.
The underlying model is OpenAI's Whisper, trained on 680,000 hours of multilingual audio according to the guide. That scale explains the tiering: languages with more real-world audio in the training set transcribe more reliably, and the guide is explicit that fast speech, poor audio quality, technical jargon, and mumbling hurt accuracy more than accent does.
The service also detects language automatically, including mid-recording switches, and outputs non-Latin scripts directly — Chinese characters, Arabic script, Cyrillic — rather than forcing a transliteration.
Source: BrassTranscripts' non-English transcription guide.
Why This Is a Content Problem, Not an Audio Problem
A recording sitting on a hard drive doesn't rank for anything. Turned into text, it becomes a page Google can index and a searcher can find.
Most businesses already have raw material in more languages than they realize. A sales call where a prospect answered in Spanish. A webinar with an international attendee asking questions in the chat, answered live in a second language. A podcast guest interview conducted partly in another language because that's what felt natural to the guest. None of that gets published, because nobody thinks of a recorded conversation as unpublished content — it's just a call that happened.
Transcription is the step that converts a recording you already made into something searchable. BrassTranscripts charges a flat $2.50 for files up to 15 minutes and $6.00 for files from 16 minutes up to two hours, with no subscription and no account required, and speaker identification comes included rather than as an upsell. Run the recording through, get text back, and you have raw material for a blog post, show notes, or video captions in whatever language the recording was actually in.
Finding the Signal in Your Own GSC Data
Ask Brass-SEO to pull your Search Console data broken down by query and country together — that combination is what surfaces a language gap, because a query alone or a country alone rarely tells the full story.
What you're looking for: a country where your dominant search language clearly isn't your site's language, showing meaningful impressions on a query, positioned somewhere in the middle of the results page rather than buried past page three. Position in the 5-to-15 range is often the sweet spot — high enough that Google sees some relevance, low enough that dedicated content would likely move it. Cross-check that against Google Analytics 4's country breakdown for the same period. If a country shows impressions in GSC but near-zero engaged sessions in GA4, that's a page ranking on incidental relevance, not content built for that audience.
Both GSC and GA4 are required to see the full picture. GSC alone won't tell you whether the traffic that does arrive actually sticks around.
None of this requires high traffic volume to be useful. A handful of impressions on a handful of queries, repeated across a few sessions, is enough to establish that a pattern exists rather than a fluke.
From Signal to Published Content
Once you've confirmed the gap, you have two ways to fill it: transcribe a recording you already have that touches the topic, or record something new specifically for that audience.
The existing-recording path is usually faster. Search your own archive of calls, webinars, and interviews for anything touching the topic where the gap appeared, even if the recording is mostly in your primary language with a segment in the target one. Transcribe it, and if BrassTranscripts' automatic language detection catches a mid-recording switch, you get both languages cleanly separated in the output rather than mashed into one transcript. From there, shape the relevant portion into a standalone post, add the context a searcher in that country would need, and publish. The new-recording path makes sense when no existing audio covers the topic — record a short answer to the actual query, in the target language if you or someone on your team can produce it, and transcribe that instead.
Either way, the loop closes back where it started: publish the page, then check GSC again in a few weeks to see whether the query that flagged the gap starts converting impressions into clicks.
See It Happen for One Real Business
The workflow above played out for one real Brass-SEO customer, with the actual GSC positions, GA4 engagement numbers, and published content to show for it.
Read One Customer. Two Posts. How Brass-SEO Found the Gap. for the full walkthrough — a real country-by-query analysis, run in a single session, closed with content published on BrassTranscripts' blog.
What This Does Not Solve
Tier 3 language accuracy still needs human review before anything gets published — the guide is direct about that, and skipping the review step on a language with limited training data risks publishing a page with real errors in it.
Finding the signal also depends on having GSC and GA4 connected and enough historical data to see a pattern, not a single day's noise. And transcription only produces raw material; someone still needs to shape a transcript into a page that actually reads well and answers a search intent, rather than pasting the raw text and calling it done. The tools remove the two biggest blockers — seeing the data and producing the base content — but the editorial judgment in between still needs a person.
Check Your Own GSC Data for Language Gaps
Connect Brass-SEO to Google Search Console and GA4 and ask it to break your query data down by country. Start your 3-day free trial →
Frequently Asked Questions
How Do I Check If I Have Search Demand in a Language I Don't Serve?
Connect Google Search Console and Google Analytics 4 to Brass-SEO, then ask it to break your query performance down by country. Look for countries where the primary language differs from your site's language but you're still generating impressions on relevant queries. Cross-reference against GA4's country data — impressions with near-zero engagement usually mean the traffic is arriving on incidental relevance rather than content built for that audience.
Is BrassTranscripts Accurate for Every One of the 99+ Languages It Supports?
No, and its own guide says so directly. Accuracy is tiered: professional-grade for English, Spanish, German, French, Italian, Portuguese, and Japanese; good for Dutch, Polish, Russian, Korean, Mandarin, Arabic, Hindi, and Turkish; and variable for the remaining 80+ languages, where the guide recommends human review before publishing anything transcribed.
What Does BrassTranscripts Cost?
$2.50 flat for files up to 15 minutes, and $6.00 flat for files from 16 minutes up to two hours. There's no subscription and no account required, and automatic speaker identification is included in both prices.
Is Brass-SEO the Same Company As BrassTranscripts?
Both are built by Copper Sun Content and Creative, LLC, but they're separate products with separate signups. Brass-SEO is $25/month for AI-powered SEO analysis of your GSC and GA4 data. BrassTranscripts is pay-per-file transcription with no subscription. Neither requires the other, but Brass-SEO's data can point to a content gap that BrassTranscripts is one practical way to fill.
Do I Need a Recording Already in the Target Language to Use This Workflow?
Not necessarily. If you have any recording that touches the topic, even mostly in your primary language, BrassTranscripts' automatic language detection separates any mid-recording language switches in the transcript, which you can shape into a targeted page. If nothing existing covers the topic, recording something new and transcribing that works just as well.