Despeech

About Despeech

Despeech is a Whisper-based audio transcription API built around one specific convenience: it accepts a URL rather than an uploaded file. A single POST with an audio URL causes Despeech to download the file, queue it, transcribe it and store the result, which removes the download-buffer-upload plumbing that a developer would otherwise have to build and host in order to transcribe audio that lives somewhere else — a podcast feed, a CDN, a recording platform. That framing, ship features not infrastructure, is what separates it from the raw model APIs it competes with. Transcription runs on Whisper with speaker diarization and support for more than ninety languages, and the audio file itself can be downloaded back out. The vendor publishes a direct price-per-minute comparison against OpenAI, Deepgram and ElevenLabs on its own pricing page, positioning itself as the cheapest of the four at ongoing rates while noting that the comparison excludes promotional offers — an unusually specific and checkable claim for a small API vendor. Every plan, including the free trial, carries the same feature set: download, scheduling, transcription, diarization, ninety-plus languages and full API access. The tiers differ only in included hours and the marginal rate for hours beyond them, which makes the pricing legible in a way credit systems usually are not.

Key Features

Tags