Parse split result error: invalid character 'A' after top-level value

[WingHaa]

Using fasterwhisper and OpenAI/Gemma with LM Studio. as LLM and getting error when trying to parse split text output from the LLM.
Am I missing some configuration? The LLM is returning string instead of json as prompt asked for.
Version: Latest release.

{"level":"info","ts":"2025-07-05T13:30:42.283+0700","caller":"fasterwhisper/transcription.go:26","msg":"FastwhisperProcessor转录开始","cmd":"./bin/faster-whisper/Faster-Whisper-XXL/faster-whisper-xxl.exe --model_dir ./models/ --model large-v2 --one_word 2 --output_format json --language zh --output_dir tasks\\06_5）转身（1）_mp4___EFVj tasks\\06_5）转身（1）_mp4___EFVj\\split_audio_001.mp3"}
{"level":"info","ts":"2025-07-05T13:32:00.769+0700","caller":"fasterwhisper/transcription.go:32","msg":"FastwhisperProcessor转录json生成完毕","audio file":"tasks\\06_5）转身（1）_mp4___EFVj\\split_audio_001.mp3"}
{"level":"info","ts":"2025-07-05T13:32:00.781+0700","caller":"fasterwhisper/transcription.go:84","msg":"FastwhisperProcessor转录成功"}
{"level":"info","ts":"2025-07-05T13:32:00.815+0700","caller":"service/audio2subtitle.go:278","msg":"Transcribe completed","taskId":"06_5）转身（1）_mp4___EFVj","splitId":1}
{"level":"info","ts":"2025-07-05T13:32:16.025+0700","caller":"service/audio2subtitle.go:315","msg":"Translate completed","taskId":"06_5）转身（1）_mp4___EFVj","splitId":2}
{"level":"info","ts":"2025-07-05T13:32:16.026+0700","caller":"service/audio2subtitle.go:1177","msg":"splitTranslateItem long sentence detected, need split","item":{"OriginText":"我先铺垫,让你们明白有这个腰的存在","TranslatedText":"Let me lay the groundwork first to help you understand the existence of this waist."}}
{"level":"error","ts":"2025-07-05T13:32:19.770+0700","caller":"service/audio2subtitle.go:1205","msg":"splitLongSentence parse split result error","error":"invalid character 'Y' after top-level value","response":"1\nYou often hear 'footwork, footwork, footwork,\n经常你们会听到转身转身转身,\n\n2\nso what’s the difference between footwork and swinging?\n到底转身和挥拍有什么区别呢?\n\n"}
{"level":"error","ts":"2025-07-05T13:32:19.771+0700","caller":"service/audio2subtitle.go:1180","msg":"splitTranslateItem splitLongSentence error","error":"parse split result error: invalid character 'Y' after top-level value","item":{"OriginText":"经常你们会听到转身转身转身,到底转身和挥拍有什么区别呢?","TranslatedText":"You often hear 'footwork, footwork, footwork,' so what’s the difference between footwork and swinging?"}}
{"level":"error","ts":"2025-07-05T13:32:19.771+0700","caller":"service/audio2subtitle.go:320","msg":"audioToSubtitle audioToSrt splitTranslateItem err","taskId":"06_5）转身（1）_mp4___EFVj","splitId":0,"error":"split long sentence error: parse split result error: invalid character 'Y' after top-level value"}
{"level":"info","ts":"2025-07-05T13:32:19.771+0700","caller":"service/audio2subtitle.go:304","msg":"Begin to translate","taskId":"06_5）转身（1）_mp4___EFVj","splitId":5}

Minimal config.toml

[llm]
    base_url = "http://localhost:1234/v1"
    api_key = ""
    model = "gemma-3-12b-it"
    json = true


[transcribe]
    provider = "fasterwhisper"
    [transcribe.fasterwhisper]
        model = "large-v2" # fasterwhisper的本地模型可选值：tiny,medium,large-v2。建议medium及以上

[puji4810]

Please change json = true to json = false in your config.toml. This should resolve the parsing error.

[WingHaa]

Thank you. Solved.