There are a lot of transcription services out there. They provide varying levels of accuracy and value, and have varying relevancy for various things you’d want to transcribe. 

The best transcription services help you easily transcribe audio and video into text that’s instantly legible and usable. Some transcription services, like Descript, use AI for faster transcription; some use humans for higher accuracy (Descript does this too, if you use White Glove Transcription). 

Of course, you can transcribe something yourself — it’s probably your cheapest option, though also probably the slowest, and the most horrible. You will need to listen to the audio file and type what you hear, which will likely mean listening to every part of your recording a few times to transcribe it accurately.

To help you choose a transcription service based on your needs, we’ve created a list of some of the most prominent options. Obviously, we think Descript is the best, and we think you’ll agree if you try it. But we’re not going to pretend the others don’t exist, and we want you to make an informed decision.

Understanding speech-to-text software

Speech-to-text software, or automatic transcription, typically uses artificial intelligence technology to detect what someone is saying — the individual words within speech — in either an audio or video file, and then creating a written version of that audio. These automated services deliver a transcript faster and at a much lower cost. However, they’re sometimes not as accurate as human transcription.

When using a speech-to-text program, you upload an audio or video file to an application or website which will process your audio using automatic speech recognition (ASR). Within minutes, typically, you have a transcript.

Some automatic speech-to-text software has extras, such as speaker identification (where the program identifies who is speaking and assigns a label to each person) and time stamping. Some services may even allow you to edit the transcript in their app, saving you time when cleaning up the text. Some also offer upgrades or higher tiers where a human transcriptionist will look over your transcript from video or audio and make any changes to ensure accuracy. Descript does all of these things. 

What to consider when choosing a transcription service

As you’re doing your research on the best transcription service for you, consider the following factors: 

• Cost. In most cases, you'll be charged on a per-minute basis, whether you're using automated or manual transcription services. Some companies have a subscription-based pricing model for those who need things transcribed frequently.  
• Turnaround time. The fastest transcription will certainly be an automatic transcription service using speech-to-text software. Human transcriptionists may be more accurate, but you may have to pay extra to get your transcript faster (and even then, it might take a few hours). 
• Accuracy and quality. Human transcription is going to be the most accurate, though humans do frequently make mistakes so don’t expect absolute perfection. Human transcribers can also identify speakers and work with poor audio quality, accents, or audio that is hard to understand. Speech-to-text software can deliver up to 95% accuracy and, in the case of Descript, identify speakers. It generally works best with high-quality audio recordings (the software makes more mistakes when the audio is hard to hear). ‍
• Editing features. Many services have built-in editing software, allowing you to make changes to correct any transcription errors before exporting it to different file types. Look for editing interfaces that are easy for you to use, allowing you to choose between playback speeds, rewind when necessary, and highlight parts of the text to sync with the recording.

The 9 best transcription services

Again, we’re going to tell you Descript is the best transcription service, because we work here — but also because we believe it. We also don't expect you to take our word for it. Here's a breakdown of some of the leading options that are out there, and their capabilities (note that pricing is as of January 2022):

1. Descript 

• Human transcription: Yes
• Automated transcription: Yes
• Editor: Yes

Descript’s transcription service is powered by the most accurate automated transcription services out there; we routinely run tests to determine which transcription engine is the most accurate, and then plug the winner into our app. 

If you need a level of accuracy no robot can provide, Descript’s human-powered White Glove Transcription will get you up to 99% accuracy. 

Descript’s automatic speaker detection lets you tell it how many speakers are on your audio and add their names. Then as it transcribes it assigns those names to the appropriate audio. When the software gets it wrong, you can fix it in an eye blink as you read. 

Because AI-powered transcription isn’t perfect – nor is humankind – Descript makes it incredibly easy to correct your transcript using a few keyboard shortcuts. You can fly through your transcript correcting words, punctuation, and capitalization using just one hand. The Correction Wizard makes it even faster by identifying likely mistakes. And you can control playback speed with a simple toggle or keystroke, to make the audio move at the speed you need for steady transcript correction.

Descript’s most unique and powerful feature is the ability to edit the underlying audio or video itself by editing the transcript. For instance, if you want to take out parts of the conversation, all you need to do is delete the text in the transcript. Same with cutting and pasting, or clipping to a new composition. The end result is you can focus on your content, not on the drudgery of editing waveforms.

2. Scribie

• Human transcription: No
• Automated transcription: Yes
• Online editor: Yes

Scribie offers both automated and manual services with speaker detection and, according to the company, 99% accuracy. Automated transcriptions start at $0.10 per minute and have a turnaround time of 30 minutes. These transcripts have an 80% to 95% accuracy — the accuracy of your transcription will depend on your audio quality, such as background noise and cross-talking between speakers. You can edit your transcription using their online editor, so you can edit the text while listening to the audio.

3. Rev

• Human transcription: Yes
• Automated transcription: Yes
• Online editor: Yes

Rev is a leader in a wide variety of transcription services, including the ability to turn those transcripts into English and foreign language subtitles. Rev can complete automated transcriptions within a matter of minutes, and their manual services have a turnaround time of 12 hours. However, their web editor doesn’t offer many features or functionality and the price can increase quickly, especially if you need to transcribe a lot of audio or translate a lot of captions. 

4. Temi

• Human transcription: No
• Automated transcription: Yes
• Online editor: Yes

Temi works with both audio-only and video files using what the company calls their advanced speech recognition software. Each transcription includes access to their web editing software, speaker identification, and custom timestamps. Temi’s automated transcripts offer 90% to 95% accuracy and have a turnaround time of minutes. Temi also offers a free trial for up to 45 minutes of transcription.

5. Otter

• Human transcription: No
• Automated transcription: Yes
• Online editor: Yes

Otter only offers AI transcription services with turnaround times in minutes. Users can sign up for the free tier, which offers transcriptions for up to 30 minutes, and 600 minutes per month in total. You can also sign up for a subscription starting at $8.33 per month. Their software allows you to edit and annotate transcripts, and easily search and organize your documents across desktop and mobile devices. 

6. TranscribeMe

• Human transcription: Yes
• Automated transcription: Yes
• Online editor: No

You’ll get pretty fast turnaround times with TranscribeMe — most customers receive their transcripts by the next business day. Their automated transcripts have a 98% accuracy, though if you want more accuracy you can have a human review the transcript (turnaround times will be longer). TranscribeMe also offers HIPAA-compliant transcriptions (for the medical industry) and legal transcriptions (where the text needs to be extremely accurate). If you need to, the company offers transcription in languages such as Spanish, Mandarin Chinese, German, French, and Japanese.

7. Transcription Panda

• Human transcription: Yes
• Automated transcription: No
• Online editor: No

Transcription Panda offers both automated and human transcription services. You can upload your audio in a variety of formats, such as MP3, M4A, WMV, AIF, AMR, WMA, OGG, WAV, or links from YouTube videos. However, their “rough draft” transcription—their basic service—is higher compared to most of the others on this list because it uses human transcribers. Transcribe Panda only delivers your transcript via Microsoft Word, which may be a disadvantage if you prefer editing your text in other file formats. 

8. GoTranscript

• Human transcription: Yes
• Automated transcription: Yes
• Online editor: Yes

GoTranscript offers human transcription services and the potential for fast turnaround times (as little as six hours). The company also claims customers will get legal-level accuracy. The price you pay will depend on turnaround times, whether you need timestamps, and if you have more than one speaker. Go Transcript also offers a customer loyalty discount—the more minutes you get transcribed, the more you'll save. 

9. GMR Transcription

• Human transcription: Yes
• Automated transcription: No
• Online editor: No

This company offers human transcription services using US-based transcriptionists with a wide range of turnaround times to suit your budget. There's also the ability to translate to other languages (Arabic, Mandarin Chinese, and Spanish). GMR can handle difficult audio with high accuracy transcription because they use human transcribers. Unfortunately, the pricing is higher compared to most competitors—you pay more for more speakers and for adding timestamps, which many other companies include for free.