# Speech to text options

There are three options for speech to text APIs that you can use with this system.

Check them out individually for extra setup instruction.

{% hint style="danger" %}
Unfortunately **Speechmatics** is no longer an option as they discontinuing their web portal (meaning the access to STT with API key will no longer be supported) from 30th of Sept 2020. See their [deprecation notice for more details](https://www.speechmatics.com/transcription-web-portal-deprecation-notice).
{% endhint %}

1. [~~Speechmatics~~](https://autoedit.gitbook.io/autoedit-3-user-manual/speech-to-text/speechmatics)
2. [AssemblyAI](https://autoedit.gitbook.io/autoedit-3-user-manual/speech-to-text/assemblyai)
3. [Mozilla Deep Speech](https://autoedit.gitbook.io/autoedit-3-user-manual/speech-to-text/mozilla-deepspeech) (Open Source, need to download 1.8gig for the English language model)
4. [Pocketsphinx](https://autoedit.gitbook.io/autoedit-3-user-manual/speech-to-text/pocketsphinx) (Open Source, integrate inside the app, no extra setup needed)

## Overview

### AssemblyAI  <a href="#ibm-watson-stt-service" id="ibm-watson-stt-service"></a>

* Free tier: 5 hours free per month
* Competitive [pricing at $0.0003 per second](https://www.assemblyai.com/pricing)
* Easy to setup credentials&#x20;
* Generally very accurate (my opinion, judge for yourself)
* For now only support for English but more coming soon

### ~~Speechmatics~~  <a href="#ibm-watson-stt-service" id="ibm-watson-stt-service"></a>

* ~~1 hour free credit with new account~~&#x20;
* ~~Easy to setup credentials~~&#x20;
* ~~Generally pretty accurate (my opinion, judge for yourself)~~
* [~~28 languages, see full list~~](https://www.speechmatics.com/language-support/)&#x20;
* ~~Including support for "accent agnostic global english".~~

### Mozilla Deepspeech  <a href="#gentle-open-source-stt" id="gentle-open-source-stt"></a>

#### Pros:

* Free as in free speech as well as in free beer.
* Working locally on your machine. No internet connection needed because of that, good for sensitive material.
* Open source
* requires to download 1.8gig for the English language model

#### Cons:

* Not as accurate as commercial STT but still pretty good (in my opinion, but decide for yourself).
* Only support US english STT.

[Check out for extra setup instructions for Mozilla Deepspeech ](https://autoedit.gitbook.io/autoedit-3-user-manual/speech-to-text/mozilla-deepspeech).

### Pocketsphinx  <a href="#pocketsphinx-open-source-stt" id="pocketsphinx-open-source-stt"></a>

#### Pros:

* Free as in free speech as well as in free beer.
* working locally on your machine. no internet connection needed because of that, good for sensitive material.
* [Open source module in autoEdit ](https://github.com/OpenNewsLabs/autoEdit_2)originally extracted from [Videogrep](https://github.com/antiboredom/videogrep) project.

#### Cons:

* Not as accurate as IBM or Gentle.
* Only support US english STT.
* Transcription takes a little longer then the length of the media. (eg 27 min takes 30 min to transcribe).

[Pocketsphinx does not require extra setup to use.](https://autoedit.gitbook.io/autoedit-3-user-manual/speech-to-text/pocketsphinx)


---

# Agent Instructions: Querying This Documentation

If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter:

```
GET https://autoedit.gitbook.io/autoedit-3-user-manual/speech-to-text.md?ask=<question>
```

The question should be specific, self-contained, and written in natural language.
The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.