Concatenate/Combine multiple audio files into 1

[Traviskn]

I've been loving the new rotateIntervalBytes option in the AudioRecorderFileOptions for building crash-resilient recordings! I keep the intervals pretty short to only a few seconds to minimize potential audio loss to only the last few seconds if there is a crash. For long recordings I end up with many small files, and it would be really convenient if there was an API available to combine them back into a single audio file to upload to my server. Since this library seems to embed ffmpeg I wonder if that could be leveraged under the hood

[Traviskn]

Having codex take a look at what this could look like:

The cleanest approach here is a new standalone async API, not work inside AudioRecorder.stop():

 concatAudioFiles(inputPaths: string[], outputPath: string): Promise<string>

 stop() already returns ordered segment paths as FileInfo.paths, including rotated files: packages/react-native-audio-api/src/types.ts:103, packages/react-native-audio-api/common/cpp/audioapi/HostObjects/inputs/AudioRecorderHostObject.cpp:63. The native recorders
 collect those segment paths during rotation on both platforms: packages/react-native-audio-api/android/src/main/cpp/audioapi/android/core/AndroidAudioRecorder.cpp:329, packages/react-native-audio-api/ios/audioapi/ios/core/IOSAudioRecorder.mm:412.

 Implementation-wise, I would not embed an FFmpeg CLI command. This repo links libavformat/libavcodec directly, and Android already has an FFmpeg writer built around avformat_alloc_output_context2, avformat_write_header, av_interleaved_write_frame, and
 av_write_trailer: packages/react-native-audio-api/android/src/main/cpp/audioapi/android/core/utils/ffmpegBackend/FFmpegFileWriter.cpp:156. So the best fit is a common C++ AudioFileConcatenator using libavformat packet remuxing:

 1. Normalize file:// paths to filesystem paths.
 2. Open each input with avformat_open_input.
 3. Find the audio stream and validate same codec/sample rate/channel layout/extradata.
 4. Create one output context and stream from the first input’s codec params.
 5. For each input packet, rescale timestamps into output time base and add a cumulative timestamp offset.
 6. Write packets with av_interleaved_write_frame.
 7. Write trailer and resolve the output path.

 That gives fast, no-quality-loss concat for same-format rotated files. If validation fails, return a clear error rather than silently transcoding.

 One caveat: for rotated m4a/AAC, each segment is independently encoded, so packet-copy concat may preserve tiny encoder-delay/padding boundaries.
 If truly gapless output matters for AAC, add a slower fallback mode that decodes all segments and encodes once. For long recordings, I’d start with packet remuxing as the default and document the AAC caveat.

[Traviskn]

I put together a PR that I have tested locally in my app on both iOS and Android to concatenate audio files together at the end of a recording session: #1045

[TaylorWu21]

I am running into the same issue. Would love to see this shipped!

[mdydek]

@TaylorWu21 @Traviskn landed, should be available in today's nightly build

[Traviskn]

Awesome! Thanks for all your help getting that PR over the line