I use a LOT of transcription services. We are a research company and record and transcribe interviews in bulk.

However, getting a service that not only provides API for high-quality transcription, but also speaker identification, and sentence breakdowns has proven tricky. And getting one as a LTD was basically an impossibility.

Until Salad.

First up, let me state that Salad is NOT a turn-key solution for uploading and transcribing. If you want that, then Otter.ai or Transcript LoL may be options (among many others).

However, if you want to build powerful transcription workflows into your existing tech ecosystem, then Salad is the best I've found. You WILL need to have some programming experience, but the automations you can achieve are vastely superior to stand-alone products. I found that with make.com, I was able to get a 'no-code' workflow running with Salad, but that still required me to understand how to develope API calls and pass informatiojn programatically.

Also, note that the Salad transcription is returned as a JSON script, with a LOT (and I mean a LOT!) of information you do not get from 'out of the box' transcription services. This means you will need to code to extract just the information you need for each use case you have. Again, that means a little bit of coding.

Finally, you do not 'upload' audio files to Salad. Rather, you upload them to a server, and pass the URL of the file to the Salad API. This means you'll need a web server or similar service to manage the audio. I found pCloud worked just fine. However, many file sharing services (dropbox, google drive) will not work, because they push all sort of HTML code in front of the audio file. So a little set up is needed regarding how you pass audio to the Salad API.

In short, Salad Transcription API is a backend service to power YOUR OWN solutions. It is a developer-oriented service. If you are not a code nerd or do not have access to a coder, then this product is probably not for you.

But if you have the need to transcribe-enable all your workflows... then I've not found a better value or higher quality product. And I have searched a lot.

Example use cases

1. Our CRM (Zoho) allows sales people and researchers to record voice notes on contacts, accounts, leads and opportunities. With Salad and Zoho Flow (or make.com, zappier, etc) I can grab the audio note, transcribe it, run it through an AI to tidy up the text,and add it to the orginal note in the CRM. That sounds like a small feature, but it dramatically improves the infdormation we keep (and can search) in the CRM, and makes our AI-power meeting preparation notes far more accurate and meaningful.

2. Have create survey forms that allow people to record responses verbally. With Salad, we can process those responses into text, then pass that into AIs for sentiment and tag extraction (classification). I created a test of this workflow and it will work perfectly.

3. Our researchers can record interviews and meetings , starting by asking each person in the forum to provide their name. The recordings are transcribed by Salad, which then allocates names to the speakers, and creates a table of the discussion, with each speakers responses aligned to each other. Even with products like Otter.ai, aligning coversations was a manual task.

4. A fun project (I am still working on). Takes a recording of a tabletop role-playing game, determins the speakers and their characters, uses AI to reframe the game adventure as a narrative (story) summary of the adventure. Yeah. That one is going to burn a lot of AI credits. :-)

Summary:

If you are looking at transcription that can identify speakers and perovide outputs in a programatic form, Salad is a must-have deal.

I purchased tier 5, which is roughtly twice the amount of recording time we currently use. I want ht headroom as with this tool, I can do a whole lot more with transcription.

In short, a must-buy for developer-minded folks.