FAQs

Axon Auto-Transcribe, with the transcript on the evidence detail page and Transcription Assistant, uses artificial intelligence (AI) to accelerate the review and transcription of evidence, creating significant time and cost savings.

This article contains frequently asked questions about Axon Auto-Transcribe. Select a link in the In This Topic menu to jump directly to a question.

Availability and access

Where is Auto-Transcribe available?

The United States (US3/Federal and US4/Enterprise), Canada, UK, and Australia

Who can view the Auto-Transcripts?

Any Pro license user who can view an evidence item can also view the associated Auto-Transcript. Basic users are limited to downloading the Auto-Transcript as a Word document only. This limitation also applies to any shared evidence.

See Auto-Transcribe licensing for more information.

Can I control who can Auto-Transcribe within my organization?

Yes, there is a permission for Auto-Transcribe. It can be set to either any evidence, my group's evidence, my evidence, or prohibited.

See Auto-Transcribe permissions for more information.

Accuracy and speech recognition

How accurate is it?

We chose the best available speech recognition solution for public safety data, not only for accuracy but also for equitable performance among diverse populations. We use the highest industry standard for measuring accuracy—every letter of the word must be correct for us to get credit. On public safety data, we average 80% accuracy, reaching levels in the high 90% range as the audio gets clearer. Microphone quality and recording environment are the two biggest factors in Auto-Transcript accuracy.

Our testing includes variations of English, encompassing several different local accents and assorted styles of body camera and interview room video. Transcription Assistant makes it incredibly easy for authorized users to make corrections.

Where is the speech recognition done?

All of Auto-Transcribe's speech recognition is completed in the Axon data center in the same country where the recording was made.

Will the speech recognition get better over time?

Yes. We release updates every few months. To receive release notes, contact your Axon representative.

What if the Auto-Transcript is challenged in court?

All transcripts, whether they are machine-generated or not, start with an Unverified status. A user with permission to mark transcripts as verified must explicitly sign off on the transcript's accuracy. At that point, the transcript becomes read-only (uneditable). This transition is logged in the Evidence Audit Trail. If mistakes are found, the transcript can be marked Unverified again in order to make changes.

See Transcript verification for more information.

Language support

Which languages are supported?

Auto-Transcripts and translations of Auto-Transcripts are fully supported for the following languages:

Dutch (Netherlands)
English (United States)
French (Canada)
French (France)
German (Germany)
Italian (Italy)
Portuguese (Brazil)
Spanish (Spain, United States)
Multilingual English-Spanish
- If your organization serves a multilingual population, it is highly recommended that you set your default language to Multilingual English-Spanish to ensure that your Auto-Transcripts are generated as expected.

Over 50 additional languages are available through a 'Public Beta' program. To request that a beta language be activated for your agency, contact your CSM or Axon Sales Representative.

For more information, see Language support.

What happens if a file contains multiple languages?

The Multilingual English-Spanish transcription model is generally available. If your organization serves a multilingual English-Spanish population, it is highly recommended that you set your default language to Multilingual English-Spanish to ensure that your Auto-Transcripts are generated as expected.

The Multilingual English-French model is available through a 'Public Beta' program.

If your transcript contains languages other than those listed above, we recommend that you download the primary language transcript as a Word document then re-transcribe the file in the other language. You can then copy from your Word document into the second language transcript to merge the two transcripts.

How does it work?

Automatically transcribed files (e.g., body camera and Axon Capture recordings when on the Unlimited plan), appear in the default language defined by the organization's administrator.

For files not automatically transcribed, you can select which language the transcript should be transcribed into.

Do additional languages cost extra?

Nope! They're included in the existing Auto-Transcribe licenses and minutes.

Does bulk transcribe overwrite existing transcripts?

Nope! If you bulk transcribe multiple files with a specific language, any file that has an existing transcript (regardless of its language) will be skipped.

What about translation?

Translations can be generated for audio and video evidence files.

Auto-Transcripts and translations of Auto-Transcripts are fully supported for the following languages:

Dutch (Netherlands)
English (United States)
French (Canada)
French (France)
German (Germany)
Italian (Italy)
Portuguese (Brazil)
Spanish (Spain, United States)
Multilingual English-Spanish
- If your organization serves a multilingual population, it is highly recommended that you set your default language to Multilingual English-Spanish to ensure that your Auto-Transcripts are generated as expected.

Over 50 additional languages are available through a 'Public Beta' program. To request that a beta language be activated for your agency, contact your CSM or Axon Sales Representative.

When working with translations, keep the following in mind:

Translations are only available to Auto-Transcribe Unlimited customers.
Additional languages will become available once they have undergone testing and evaluation to ensure accuracy and bias prevention.
Transcripts containing both Spanish and English or "Spanglish" can be translated as long as the Multilingual English-Spanish language model is used to generate the original transcript.
Translations are AI-generated and unverified, so they may contain errors.
Translations are intended for on-demand quick evidence review. They can't be edited and verified like regular transcripts.
Translations are single-session; if you leave the evidence details page after generating a translation and return to it later, the translation will no longer appear and must be regenerated.
After generating a translation, you can download it as a DOCX file.
Translations can't be created for documents, images, or video frames.

What happens if a file is transcribed in the wrong language?

You can re-transcribe the file in another language. When re-transcribing a file, you will see a dialog warning you that any edits you made to the original transcript will be lost. This does not charge the user any additional minutes. To re-transcribe a file:

Open Transcription Assistant and select Change Language.
Choose the new language from the Transcript Language menu.
Select Start.

See Language support for more information.

Specifications

Which file types does Auto-Transcribe support?

Auto-Transcribe can be used with any audio or video file type supported by Axon Evidence.

Are there limits on the video or audio file duration?

Currently, Auto-Transcribe supports files up to 10 hours in length.

How are transcription minutes calculated?

By duration of the video/audio file.

See Auto-Transcribe licensing for more information.

Why isn't the Auto-Transcribe button showing up on the evidence detail page?

Please ensure the following:

The file type is supported by Axon Evidence and is either audio or video.
You have a pro license for Axon Evidence, and your assigned Role has been granted the Auto-Transcribe permission for the evidence.
Your organization has an Auto-Transcribe license, either as an add-on or as a la carte minutes.

Are there browser requirements for Auto-Transcribe?

Auto-Transcribe is supported on the latest stable release of Google Chrome, Mozilla Firefox, Microsoft Edge, and Safari.

Requesting and Sharing Auto-Transcripts

Once I request an Auto-Transcript, if someone else wants to look at it, do they have to request it again?

No. Once a piece of evidence has been Auto-Transcribed, anyone with the permission to view it can see the associated Auto-Transcript.

What are all the ways an organization can request Auto-Transcripts?

Auto-Transcripts can be requested from these locations:

Evidence detail page for a single piece of evidence
Evidence search page for multiple files (including within Cases)

See Request an Auto-Transcript for more information.

If I share evidence with someone else, do they have to request the Auto-Transcript again?

No. Auto-Transcripts can be shared along with evidence or case shares.

See Share Auto-Transcript for more information.

Auto-Transcripts and Redaction

If I redact audio, will that automatically redact the transcript?

If you request an Auto-Transcript on the redacted file, the redacted sections of the evidence will NOT appear in the Auto-Transcript.

Speakers

Are speakers automatically identified in the Auto-Transcript?

Yes. The initial Auto-Transcript inserts a line break in the text whenever it detects a speaker change. However, as with speech-to-text, these speaker assignments may not be accurate. Transaction Assistant lets you add new speakers and speaker sections, and rename speakers.

Speaker labels appear in Transcription Assistant and in Auto-Transcript downloads. For more information, see:

Transcript Downloads

Can the Word doc export include line numbers?

Yes. This feature can be adjusted when downloading the Auto-Transcript.

See Download transcript for more information.

Can I export the transcript as a PDF?

Not at this time.

Audit Trail

Are edits reflected in the audit trail?

You can see in the evidence audit trail if an Auto-Transcript was edited, including when/who edited it. Currently, you can't see a full revision history that shows exactly what edit was made.

See Auto-Transcribe audit trail for more information.

Foot Pedal Support

Does the foot pedal software work with Olympus pedals?

Ours doesn't seem to, but there is a workaround. Install the Olympus configuration tool and select the new template option on the top left. (Don't override the preset template). After you save new template, you can edit the shortcut keys. [ for rewind, ] for play, \ for fast forward.

See Foot pedal support for more information.

Keyword search across transcripts

Can my organization use keyword search?

This feature can only be used by organizations who have the Auto-Transcribe Unlimited Add-On plan. If your organization is not currently using this plan and you are interested in upgrading, contact your Axon representative.

How does keyword search work?

For instructions on using keyword search, see Keyword search across transcripts

Are additional permissions required to use keyword search?

In addition to having the Auto-Transcribe Unlimited Add-On plan, users must have the Search across Transcripts permission to use this feature. When this feature is launched in early 2024, this permission will be auto-enabled for the following Roles:

Default Admin Role
Default Investigator Role
All Roles who currently have the Edit Auto-Transcript permission

For more information, see Auto-Transcribe permissions.

What if my organization has old evidence that hasn't been transcribed?

When an organization purchases the Auto-Transcribe Unlimited Add-On plan, only new evidence is transcribed. Evidence created before the plan goes into effect is not transcribed. If you want to back fill transcripts for all of your organization’s evidence, contact your Axon representative. A cost is associated with this service, depending on the amount of evidence that needs backfill transcripts.

How does keyword search work with multiple languages?

Auto-Transcripts are available in English, Spanish, and French-Canadian, based on the language settings in your user profile. As a result, your organization may have transcripts in multiple languages. Keyword search matches text in a transcript, so while it works for other languages, it is currently optimized for English.

There is a list of words that are skipped during the keyword search, such as “a”, “an”, “the”, etc. You can still enter these words the Transcription filter, but these words are ignored when determining search result relevancy. A similar list of words currently does NOT exist for other languages (e.g., “la” for Spanish).

Keyword search does support special characters, such as “ñ”.

Axon's AI-powered features use artificial intelligence to provide insights, recommendations, or analysis based on available data and user inputs. These outputs may include inaccuracies, limitations, or outdated information. These tools are designed to support—not replace—professional judgment, legal interpretation, and human oversight. Users are responsible for carefully reviewing and verifying AI-powered content before acting on it. Axon continually works to enhance AI accuracy, mitigate bias, and uphold ethical standards. Learn more at Axon AI.