How to Use AudioPen to Draft Blog Posts From Voice Notes
Drafting by voice is faster than typing for most people — roughly 150 spoken words per minute against 40 typed. The bottleneck is turning that ramble into something readable.
AudioPen shortens that gap. It cleans the recording, restructures it, and gives you a chunk of text you can expand rather than transcribe. Here is the workflow that works for posts between 800 and 1,500 words.
What does the recording need to contain?
Not a finished article. Three things:
- The claim you want to make, in one sentence
- Two or three pieces of evidence, each with a number
- Who the post is not for
The third item is the one people skip, and it is what makes a draft sound like a person instead of a product page.
Speak for four to six minutes. Beyond that, the processing output gets vaguer — the tool compresses hard, and long recordings lose specifics.
How do you go from recording to outline?
- Record the ramble as described above.
- Pick the outline style. You get a bulleted structure, usually five to seven points.
- Re-process the same recording with the notes style. Now you have two views: structure and prose.
- Merge them by hand. The outline tells you the order, the prose view tells you the argument.
- Fill the gaps. Generated outlines are always missing the concrete detail — add your numbers, prices, version strings.
That merge step takes ten minutes and is where the article actually appears. Do not skip it.
How much of the output can you keep?
For a 1,000-word post: expect 150-250 usable words straight from AudioPen — usually the opening assertion and one strong middle paragraph.
The rest is you expanding. That is normal. The tool is a structuring aid, not a ghostwriter.
What it does well:
- Removes hesitation and repetition without you noticing
- Turns spoken fragments into complete sentences
- Keeps your voice if you speak in full sentences
What it flattens:
- Specific product names and model numbers
- Anything sarcastic or dry — tone gets neutralised
- Jokes. All of them.
Where does the workflow break?
Interviews and quotes. It paraphrases. If you need a verbatim quote, transcribe separately.
Technical explainers with code. Spoken code is useless. Write those sections by hand.
Very long source material. A 20-minute interview exported through AudioPen comes out as a 300-word summary of nothing in particular. Chunk it into topic segments first.
If the post needs supporting data, pull the numbers before you record — audio-to-text cannot invent a statistic you did not say, and that is a feature. But it also means the numbers have to already exist somewhere. AnswerRocket is where I pull warehouse figures before a data-heavy post.
What is the realistic time saving?
On a 1,000-word post: typing a rough draft takes about 50 minutes. Recording plus cleanup plus generation takes about 12 minutes, and the resulting draft needs around 30 minutes of expansion.
Roughly 10-15 minutes saved, and the draft is less polished but structurally sound. For teams publishing several posts a week, that adds up to a few hours monthly.
If the content is a video script instead of a post, the same voice notes feed a script and Blizzy handles the edit once you have recorded the footage.
Quick Summary
- Record 4-6 minutes with a claim, two numbers and a “not for you” line.
- Generate both an outline and a notes version of the same recording, then merge by hand.
- Expect 150-250 usable words per 1,000-word post; the rest is expansion.
- Product names, numbers and jokes get lost in processing — fix them manually.
- Realistic saving is 10-15 minutes per post, mostly from the structuring step.
