Documentation

Docs are under active development. Email support@spuric.com with anything that is unclear.

For customers

Quickstart: rent a GPU - sign in, pick a machine on the marketplace, choose a template (PyTorch, ComfyUI, vLLM, Jupyter), click "Rent". SSH details email within 60 seconds.
Templates - prebuilt container images for common workflows. PyTorch CUDA 12.6, ComfyUI + Flux, vLLM with GLM-5.2 / Qwen / DeepSeek, Jupyter Lab, plain Ubuntu + CUDA toolkit.
Persistent storage - mount a SPUR Volume on any machine for $0.10 / GB / month. Survives container restarts and machine moves.
Inference API - OpenAI-compatible endpoints for GLM-5.2, DeepSeek-V4 (Pro and Flash), Llama 3.3 70B, Qwen3-Coder and a vision model - plus image, video, speech, transcription, embeddings and reranking - on the SPUR-owned cluster - see the full live lineup on the Models page.

Quickstart: list a machine - apply to host, complete verification, install the SPUR agent (one curl command), configure your machine in the host dashboard, go live.
SPUR Agent - single binary, manages Docker + NVIDIA container runtime, GPU isolation, image distribution, billing telemetry. Average install time 12 minutes.
Verification flow - what we need for Tier 1, 2, and 3. Public verification table shows every active host's tier and badges.
Payouts - monthly via wire / SEPA / interac e-Transfer / SPUR Token. Earnings dashboard updates every 5 minutes.

Customer API - REST + WebSocket for listing search, instance lifecycle, billing. Documented at /docs/api/customer (coming soon).
Host API - agent registration, telemetry, payout reporting. Documented at /docs/api/host (coming soon).
Inference API - OpenAI compatible. Base URL https://ai.spuric.com/v1. Auth via API key.