Omni API is a unified inference API with an OpenAI-shaped surface. Point your
existing SDK at a new baseURL, prepay credits, and generate across
modalities — no per-provider accounts, no idle GPU bills.
Open admin console Model catalog API status
Text-to-video and image-to-video. 5 or 10 s clips at 720p, billed per second of output.
Flux.1-schnell generations, up to 4 per request, billed per image.
Whisper large-v3 with word timestamps and VAD. Sync under 5 minutes, async beyond.
Kokoro voices, streaming mp3/ogg, billed per minute of generated audio.
/v1/chat/completions returns 501 until text models land.
Receive an ok_live_… API key and prepay USD-denominated credits.
Balance and usage are queryable any time.
Keep your OpenAI client — change the baseURL and the key.
Same request shapes, same error envelope.
Each job is quoted up front and credits are reserved before any GPU spins. Poll the job or receive a signed webhook.
curl https://ai.theconsultant.chat/v1/videos/generations \
-H "Authorization: Bearer ok_live_..." \
-H "Content-Type: application/json" \
-d '{ "model": "wan-2.2-t2v", "prompt": "a red panda surfing at sunset",
"duration": 5, "resolution": "720p" }'
# → 202 { "job_id": "..." } then poll GET /v1/jobs/{id} or receive your webhook| Model | Modality | Billing unit |
|---|---|---|
flux-schnell | image | per image |
whisper-large-v3 | stt | per audio minute |
kokoro | tts | per audio minute |
wan-2.2-i2v | video | per video second |
wan-2.2-t2v | video | per video second |
Live catalog (with unit prices once published): GET /v1/models — no key required.