# Add captions

## ❓ What is it?

FireCut lets you add captions to your video.

While English is fully supported, **FireCut also lets you transcribe audio in 50+ languages**; however, please do double-check the results because this is an experimental feature.

## 🤔 How does it work?

The "Add captions" feature:

* Transcribes your audio in 2 steps - **both are done automatically**
  * (1) Sentence-level (takes \~5s per min of audio) -- gives a preview of the transcript
  * (2) Word-level (takes \~5s per min of audio) -- gets accurate timings for transcript
* Lets you modify the captions in 2 different modes: 
  * **Styled captions** (if you want captions that are styled and animated within FireCut), or
  * **Text-only captions** (if you just want a captions track that you can style yourself in Premiere)
* You can also export the captions as SRT or VTT files

{% embed url="<https://www.youtube.com/watch?v=NkCEjmI3Mbk>" %}

### 1️⃣ Styled captions mode

* **When to use:** To easily get animated captions that match popular creators' styles
* **Output:** Caption images (.PNG files) inserted in a new video track in your sequence. Even though these are images, [**they can still be edited by FireCut**](#editing-styled-captions)

### 2️⃣ Text-only mode

* **When to use:** To get a text-only captions track that you can edit in Premiere
* **Output:** New captions track (requires PPro 2022+) that contains your transcript

## 💾 Saving your brand style

Once you have changed the various settings (e.g. font, text position, animation) to match your style, just click the **+ button** to save this style as a Preset. Next time, you can just click the name of the Preset to load the style again.

<div align="left"><figure><img src="https://4122560492-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FWDn6n1C7MKZBHdDcmm6f%2Fuploads%2F2fMFrwLLJFmNXzNtBULc%2Fimage.png?alt=media&#x26;token=0eb645c6-06ee-4c91-b87f-ea6863a36162" alt="" width="375"><figcaption></figcaption></figure></div>

## ✏️ Editing styled captions

Styled captions are inserted as .PNG images in your sequence. This is done by design to give you a great deal of flexibility in how the captions are styled and animated.

If you find that you'd like to edit any of these .PNG images, it is very easy to open them up again in FireCut and make any changes by:

* Selecting the caption images in the timeline
* Going to **FireCut** > **Add captions** > **Edit selected captions** (under the Transcribe button)<br>

  <div align="left"><figure><img src="https://4122560492-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FWDn6n1C7MKZBHdDcmm6f%2Fuploads%2FNeDj8XCHKeq9i5CqsFH4%2FScreenshot%202023-12-27%20at%203.56.20%20pm.png?alt=media&#x26;token=ff67bf42-8709-49c3-aa12-3ed2173c8ddc" alt="" width="209"><figcaption></figcaption></figure></div>

This opens up the selected portion of your transcript in FireCut, allowing you to make any changes (e.g.  changing the text or style). If you want to reload the full transcript instead of just the selected captions, click the "**Load full transcript**" button

![](https://4122560492-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FWDn6n1C7MKZBHdDcmm6f%2Fuploads%2FdLfAevMewko2WG0ulu9h%2FScreenshot%202023-12-27%20at%203.55.47%20pm.png?alt=media\&token=74af7193-667d-4c72-882d-70f3916a8e25)

To add the edited captions, just click "**Add styled captions to sequence**" once you're done.

## ⚙️ Settings

1. **Scope:** You can decide whether to run the operation on your **Full sequence** or only on a certain portion that you specify with **In / Out points** (you can place these in your sequence using the `I` and `O` hotkeys)
2. **Audio language:** We are constantly adding support for new languages. However, please note this is experimental and may create unpredictable results when compared with English. This is the case for every AI tool even if they don't mention it, because [training sets for AI tend to be in English](https://www.wired.com/story/chatgpt-non-english-languages-ai-revolution/). On occasion, this might cause the operation to fail -- please do let us know (<support@firecut.ai>) if this happens, so we can improve it for you!
3. **Guidance prompt:** This textbox lets you provide a writing style to the transcription model in order to guide the transcript to be more in line with what you want. You can include things like:
   1. **Brand names / Proper nouns** that may not be spelled correctly by default (e.g. type "FireCut" in order to correct "fire cut" to "FireCut" in the transcript)
   2. **Capitalisation preference**, for example, if the transcript is coming out in lower case ("the park is green") and you want to force sentence case ("The park is green."), you can write a short sentence here in your preferred capitalisation and FireCut will try to follow that guidance
   3. **Punctuation preference**, for example, if the transcript tends to miss punctuations, you can write a short sentence here with the preferred level of punctuation and the transcription will follow your guidance
4. **Translate to another language:** This setting lets you translate your captions to a different language. Be sure to set the audio language correctly to the language your original audio is in
5. **Use voice activity detection:** If you have long silences in your transcript, this can cause problems in the resulting transcript, such as repetitions of the same word or the model adding unrecognised words. Using voice activity detection (VAD) can help to eliminate this problem

## 💡 Tips for best outcomes

* **Make sure your audio does not have long silences**, and any music / sound effects are muted
* **Remove any repetitions first**, as FireCut might not transcribe multiple instances of the same text
* **Feel free to work in the timeline**, while FireCut is processing your audio -- however, changing the timing of your clips might cause the transcription to be misaligned (FireCut's transcription follows your **sequence**'s timecode, not individual clips)

## 🙋‍♂️ Common questions, issues, watch-outs

* **"The transcribed text was wrong"** -->
  * **All languages:** Please make sure the audio is clear, there are no long silences, and no repetitions. If this doesn't fix the issue, please share your audio with us at <support@firecut.ai>
  * **Non-English:** Transcription for non-English audio is currently experimental and expected to be 70-80% correct in most cases. If you're experiencing a much worse outcome, please get in touch and share your audio with us at <support@firecut.ai>
* **"The transcript has a style I don't like" (e.g. too many capital letters, everything in lowercase, missing punctuation, etc.)** -->
  * Use the "guidance prompt" setting (see [#settings](#settings "mention") above) to guide the transcript's style


---

# 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://learn.firecut.ai/features/add-captions.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.