oruk Speech API
API by oruk · rest
Speech APIs for 25-language file transcription with Spectra-2 (oruk-spectra-2), 32-locale live transcription with Realtime, vocal emotion and speaking-style analysis, and model-specific speaker diarization. Spectra-2 returns 15 emotion and 16 speaking-style scores for each clip. Spectra-2 does not label speakers. Resonance 2 preview returns 31 continuous expression scores and six signed axes, without transcription; new organizations require separate approval. Resonance 1 is the default English file model. Hosted Orukeet is an English preview included in every subscription at $0.00045 per audio minute ($0.027 per audio hour); its separate local model supports 25 languages. Optional Orukeet tasks add their published rates. Subscription models use plan minutes. Orukeet usage.cost_microusd is usage valued at task rates, not an additional charge when covered by the monthly allowance. Only excess usage is billed. Existing credit-only accounts retain their credit balance. OrukZero…
Endpoint: https://speech-api.oruk.ai/
Verification: observed. Source: crawl.
Capabilities (19)
- Service liveness
- Inference readiness
- Classify emotion and speaking style
- Return transcript emotion style segments and tagged text
- Classify multilabel emotion
- Continuous recording and word pronunciation proficiency (0–1)
- Classify emotion and delivery with Resonance-2
- Check the Resonance emotion sidecar
- Prepare a Spectra-2 session
- Prepare a direct Spectra-2 stream
- Classify multilabel speaking style
- Detect human or AI speech with OrukZero
- Transcribe audio with a supported model
- Create a single-use Orukeet streaming token
- List public models and supported tasks
- Read organization Resonance-2 access
- Request Resonance-2 access
- Check OrukZero account access
- Get reference cost metadata