I’m looking for a transcription service that handles accents, overlapping speech, and custom vocabulary for industry terms. What should I try?

For accents, overlapping speech, and custom industry vocabulary, I’d try one of these:

Best options to evaluate

1. OpenAI Whisper-based services

   • Very strong with accents and noisy audio.
   • Good general transcription quality.
   • Some platforms built on Whisper offer custom vocabulary / prompt hints.

2. AssemblyAI

   • Good at speaker diarization and accent/noisy audio.
   • Supports custom vocabulary and domain terms.
   • Worth trying if you need a developer-friendly API.

3. Deepgram

   • Often strong for real-time transcription, overlapping speech, and custom keywords.
   • Good if you need low latency or scalable API usage.

4. Google Speech-to-Text

   • Solid for many accents.
   • Offers speech adaptation for custom words/phrases.
   • Reliable enterprise option.

5. AWS Transcribe

   • Supports custom vocabulary and speaker labeling.
   • Good if you already use AWS.

If overlapping speech is a big issue

No transcription service handles overlap perfectly, but you’ll usually get the best results from:

• Deepgram
• AssemblyAI
• Whisper-based workflows
• plus speaker diarization and, if needed, post-processing

What I’d recommend

• Try Deepgram and AssemblyAI first if you want API-based services optimized for production use.
• Try Whisper-based solutions if accent robustness is your top priority.
• Try Google/AWS if you want enterprise cloud integration and custom vocabulary features.

Quick decision guide

• Best overall accuracy with accents: Whisper-based
• Best for custom vocabulary + production API: AssemblyAI / Deepgram
• Best cloud enterprise fit: Google / AWS
• Best real-time transcription: Deepgram

If you want, I can also give you:

• a comparison table
• recommendations based on budget
• or a shortlist for medical, legal, or technical terminology.