ios speech recognition Error Domain=kAFAssistantErrorDomain Code=216 "(null)"

This will prevent two errors: The above mentioned Code=216 and the 'SFSpeechAudioBufferRecognitionRequest cannot be re-used' error.

1. Stop recognition with finish not with cancel

2. Stop audio

like so:

    // stop recognition
    recognitionTask?.finish()
    recognitionTask = nil

    // stop audio
    request.endAudio()
    audioEngine.stop()
    audioEngine.inputNode.removeTap(onBus: 0) // Remove tap on bus when stopping recording.

P.S. audioEngine.inputNode seems to be no longer an optional value, therefore I used no if let construct.

I had the same problem whilst following the same (excellent) tutorial, even when using the example code on GitHub. To solve it, I had to do two things:

Firstly, add request.endAudio() at the start of the code to stop recording in the startButtonTapped action. This marks the end of the recording. I see you've already done that in your sample code.

Secondly, in the recordAndRecognizeSpeech function, when 'recognitionTask' is started, if no speech was detected then 'result' will be nil and the error case is triggered. So, I tested for result != nil before attempting to assign the result.

So, the code for those two functions looks as follows: 1. Updated startButtonTapped:

@IBAction func startButtonTapped(_ sender: UIButton) {
    if isRecording {

        request.endAudio() // Added line to mark end of recording
        audioEngine.stop()

        if let node = audioEngine.inputNode {
            node.removeTap(onBus: 0)
        }
        recognitionTask?.cancel()

        isRecording = false
        startButton.backgroundColor = UIColor.gray

    } else {

        self.recordAndRecognizeSpeech()
        isRecording = true
        startButton.backgroundColor = UIColor.red
    }
}

And 2. Update within recordAndRecognizeSpeech from the recognitionTask = ... line:

    recognitionTask = speechRecognizer?.recognitionTask(with: request, resultHandler: { (result, error) in
        if result != nil { // check to see if result is empty (i.e. no speech found)
            if let result = result {
                let bestString = result.bestTranscription.formattedString
                self.detectedTextLabel.text = bestString

                var lastString: String = ""
                for segment in result.bestTranscription.segments {
                    let indexTo = bestString.index(bestString.startIndex, offsetBy: segment.substringRange.location)
                    lastString = bestString.substring(from: indexTo)
                }
                self.checkForColoursSaid(resultString: lastString)

            } else if let error = error {
                self.sendAlert(message: "There has been a speech recognition error")
                print(error)
            }
        }

    }) 

I hope that helps you.