I have been on the lookout for transcription software, and hoped that HappyScribe would fit the bill. Unfortunately, it's just not going to cut it, which is a big shame.

I have been using it for a couple of transcriptions. The first was not great quality, but still passable, and HappyScribe did not do a good job with it. The second was much clearer and it did an okay job, but still not as good as other software out there.

For reference, I ran the same files through Descript and Temi (made by Rev), and both did a better job.

Here are my problems, other than the not so stellar transcription, with HappyScribe:-

1. It is TERRIBLE at identifying speakers. In both interviews, my speakers sound COMPLETELY different. HS was not able to pick up the differences. Descript nailed it, Temi was mostly good. Heck, HS even changes the speaker and drops to a new paragraph half way through someone talking!

2. There is no way to bulk change speaker names... not that you want want to, given the above.

3. It needs an option to remove "umms", "uuh," false starts, "you know"s, double speak, etc. Most of my time is spent removing these.

4. Talking of tha, if you do "find & replace" but leave it blank in replace, to get rid of the above, it does not work.

Here are some suggestions I have to improve HS, other than transcription quality (which is definitely not the best, like they claim):-

1. Customisable shortcuts in the editor, so we can use buttons we are used to i other editors.

2. An option to upload multiple files for a transcription - 1 for each speaker - so HS can better transcribe each person's dialogue.

3. Need more options for export - like layout (inline, column, etc), highlighted text only.

4. If making a paragraph break and putting in no speaker name, on the export, it should just leave one carriage return, not two (where one is a blank one for a speaker name). It should assume it is merely a paragraph break, not a new speaker.

5. When clicking an unknown word, alternative suggestions would be useful.

6. Upon sharing, you should be able to control what the recipient can edit (what speaker names, etc).

7. An edit video/audio mode, whereby, if you delete text, the video/audio gets cut too.

8. An option to only export audio and video that corresponds to the highlighted text, so we can make short summary versions of our videos/audio. (See Soundtrap and Descript.)

9. The ability to host calls with multiple participants and record a track for each of them, which is then transcribed. (See Soundtrap and Descript.)

Lastly, I really think HS needs to up the 2 hour limit for one code. It's simply not enough when compared to the competition, who also give you a LOT more in regards to options and abilities in their software (yes, including SRTs, which HS keep talking about). You can also use Rev.ai to do automated transcription for $0.035 per minute and edit in Word. So why use HS?

Will any of the above make it into the roadmap?