Models

SpeechWeave offers core and max transcription models. Use core by default for the best balance of speed and cost. Use max for the best quality on noisy or complex audio.

Service modes: standard is the higher-priority mode; deferred is the lower-cost, lower-priority mode at half the standard rate. OpenAI model aliases such as whisper-1 map to core.

Features: word timestamps, SRT/VTT, webhooks, sync and async delivery, and up to 250 MB per self-serve job. 100 specifiable languages and auto detect