Question: I need a solution that can transcribe speech with high accuracy and timestamping, do you know of any options?

If you need a service that can transcribe speech with high accuracy and timestamping, Gladia is a good choice. Gladia's AI transcription API uses optimized Whisper ASR technology to deliver accurate transcriptions with speaker diarization, code-switching, and word-level timestamps. It can transcribe multilingual speech-to-text and supports end-to-end security and encryption that meets EU and US privacy standards. The service is designed to be easy to integrate with different tech stacks, so it's good for content and media, virtual meetings, workspace collaboration and call centers.

Another good option is AssemblyAI, which offers a variety of AI models for speech-to-text transcription, speaker detection, sentiment analysis and more. Its highly accurate Universal-1 model is trained on 12.5 million hours of multilingual audio data and supports more than 99 languages. AssemblyAI offers integration tools to accommodate different needs and a free tier for prototyping, with pay-as-you-go pricing for production. The service is geared for companies building their own AI products and offers data security with GDPR, PCI-DSS and SOC 2 compliance.

For a service that offers high accuracy and flexibility, check out TurboScribe. It can convert unlimited audio and video files into text with 99.8% accuracy and supports more than 98 languages. TurboScribe offers unlimited transcripts with no limits or quotas, so it's good for podcasters, researchers and businesses. It also offers speaker identification and private encryption for data security.

Last, Deepgram offers a range of APIs for speech-to-text, text-to-speech and audio intelligence with high accuracy and low latency. The service supports multiple languages and offers detailed transcription data, so it's good for speech analytics and media transcription. Deepgram's pricing is flexible, with a free API playground and a range of plans for different needs.