Contents
This post is part of a paid promotion with Notta. I had been using Notta since 2023 when they asked whether I would help with PR. Since it's a service I have used for a long time, I decided I could write about it honestly from my own experience, and that is what this post is.
I started using Notta on 5 May 2023.
I've kept the subscription running since, without cancelling, for over three years.
The amounts are masked because they reflect the prices at the time I subscribed. I started on the Premium plan and moved to the Business plan on 2 May 2024.
Written like that, it might sound like I used Notta to the full from day one. I didn't.
What I had in mind at the start was:
- Could I transcribe footage I'd shot and use it while editing?
- Could I draft subtitles and captions from it?
- Could I summarise a video for a blog post or a YouTube description?
I was doing a lot of video editing then, and a lot of my time went into watching footage start to finish while organising what was in it.
If I could read the contents first, finding the scenes I needed would get easier. With luck, subtitles and descriptions would get faster too. That was the expectation I signed up with.
In 2023 it wasn't perfect yet
Honestly, back when I started, the transcription wasn't clean enough to drop straight into subtitles.
There were misrecognitions, and videos heavy with jargon needed corrections. "Transcribe it and video editing becomes fully automatic" was not the reality.
What did help, even then, was the AI summary.
Even with small errors in the phrasing, you get the shape of it:
- What is this video talking about?
- How does the discussion progress?
- Where are the important points?
Less "produces a finished script automatically", more "organises the material and shortens the time until you can start writing". That was the first value I got from Notta.
Shaping the AI summary around how you work
These days you can go past "make the transcript shorter" and build custom templates that organise it into the shape you want.
For me that means:
- Splitting the video's key points into sections
- Pulling out what to do next as a to-do list
- Organising topics that could become blog posts or teaching material
For this post, I built a template called "video reuse notes" in the actual Notta interface.
The instruction I typed was only this.
Organise the key points of the video, and summarise the to-dos and any topics usable for blog posts or teaching material in a readable form.
Pressing the AI improve button next to the field turned that short instruction into one that outputs, separately:
- The video's key points
- Decisions made
- A to-do list including owners and deadlines
- Topics usable for blog posts or teaching material, with reasons
Even if you aren't good at writing clean prompts, you can get from "I want the to-dos separated" and "make it easier to read" to something practical. That helps.
Running the AI summary over a test recording of my own, I could compare the summary and the original transcript on the same screen. Each summary item shows the timestamp in the source audio, so you can jump back to the original statement to check something.
The video used in this post is for testing. The meeting content is entirely fictional: AI-generated meeting audio turned into a video and uploaded to Notta. It has no relation to any real meeting or organisation.
That video runs about 43 minutes. Watching from the moment the upload finished to the end of the transcription, it took under two minutes.
43 minutes of video, searchable text in less than two. Even as a single test result — that's a bit much, isn't it?
In this test, transcription finished in under two minutes after the upload completed. Processing time varies with the video and the network.
Not stopping at reading the transcript, but shaping it into something you can use for the next task: that's the part that has changed most over the years I've used it.
If you're curious, looking at the feature itself is faster.
The main reason: I can upload video I've already shot
There are plenty of services that turn audio into text. But as far as I've looked, services that let you upload an already-shot video file directly are surprisingly rare.
Notta has had this for a long time, and I have leaned on it heavily. It is the main reason I can't leave.
I have recordings of interviews, lectures and seminars. Upload those to Notta and I can grasp the whole thing from the transcript and summary first, then check only the moments I need in the original video.
Turning an interview into an article. Organising the key points of a lecture. Finding the wording for subtitles and captions. The more video you handle, the bigger the gap between this and playing everything start to finish.
Turning old video back into usable information
As the pile of old footage grows, you lose track of which video said what.
So lately I've been uploading batches of old video to Notta and organising them with transcription and AI summaries.
Read the summaries, sort by theme, and rebuild what's worth rebuilding into blog posts, teaching material or short videos. For me, Notta isn't a meeting-minutes tool. It's the entrance to reusing an archive.
Couldn't you just transcribe locally?
There are transcription AIs you can run on your own machine, and I do run AI locally as a matter of course.
But that same machine is also editing video and running development work. Tying up its processing power for hours to transcribe a pile of video isn't efficient in my setup.
With Notta, once the file is uploaded the processing happens on the web. My machine stays free for something else. For a couple of files, local is fine. For handling a steady volume, the convenience has real value.
Reachable from any device
Upload video from the laptop, check the transcript and summary later on my phone. Read it back while out. For meetings, record straight from the phone app.
The data follows me across devices, and the transcript isn't trapped on one machine. That also makes it easy to keep using.
Which is why it's been over three years
Notta isn't perfect for everything.
Transcripts sometimes need correcting, and for a couple of files a free tool or a local AI may be all you need.
The reason I've kept it since 5 May 2023 is simple.
Because I can take video I've already shot and move it into transcription and an AI summary without loading up my own machine.
And because that text and that summary are the entrance to using old footage again.
I've deliberately skipped price comparisons and fine-grained differences against other services this time.
Being asked to help with this promotion was a good excuse to look again at how I use Notta.
Since I've been paying for over three years, I might as well use it properly.
So I plan to keep writing in detail about what Notta can actually do and how I organise my archive, testing it myself as I go.
Notta isn't the only thing I use hard, of course. I use plenty of other AI services and tools, and I'll keep writing honestly about what I learn from using them.
More of those experiment logs, gradually.
If you'd like to try Notta, this referral link gives you 10% off.
The link above is a referral link. If someone signs up through it, I may receive a referral fee.
The features and screens in this post are as of August 2026 and may change. For the current information, see Notta's official page.
