Gateway

Coming soon

The gateway is the OpenAI-compatible HTTP API you call: it authenticates the request, meters usage, and routes the job to a node in the network. Chat was the first surface live on it, but the same vk_ key and base URL now also reach embeddings (metered by tokens), audio speech and transcription (metered by characters or seconds), image generation and editing (metered per image), and async video generation (metered by generated seconds, via a queued → completed → content poll loop) — one endpoint surface, authenticated and routed the same way regardless of modality. This page is being written. What will live here: how a request is routed to the node mesh, the full endpoint surface the gateway exposes across every modality, and how it authenticates and meters each call.

Next steps

• Quickstart — get a vk_ key and make your first call.
• API reference — every endpoint, generated from the live gateway spec.
• Embeddings — vector generation, metered by tokens.
• Audio — speech and transcription, metered by characters or seconds.
• Images — generation and editing, metered per image.
• Video — async generation, metered by generated seconds.