Green Compute docs
Everything on Green Compute is one HTTP API at https://api.green-compute.com. Rent a GPU over SSH, billed per minute, or call open models through an OpenAI-compatible endpoint.
Using an AI agent? Point it at green-compute.com/llms.txt. Every page here is also available as Markdown by adding
.md to its URL.Quickstart: rent a GPU with the APIFrom an API key to an RTX 4090 you can SSH into, and back, in six HTTP calls.For AI agentsMachine-readable entry points, what needs a human, and rules of thumb for agents using Green Compute.GPU rental referenceWorkloads, deployments, lifecycle, SSH, ports, disk, resume and RTX 5090 notes.Create an SSH keyMake an SSH key on macOS, Linux or Windows and add its public half, so you can log in to the GPUs you rent.Inference API (OpenAI-compatible)Chat completions, streaming, tool calling and reasoning models at api.green-compute.com/v1.PricingPer-minute GPU pricing, per-token pricing for shared models, hourly pricing for your own models, and the live machine-readable price list.Adding creditCard, USDC/USDT, TAO and alpha: exact invoice amounts, verified wallets, and what to do if a payment isn't credited.Use Green Compute from AI assistants (MCP)Connect Claude or another MCP client to https://api.green-compute.com/mcp to rent, watch and stop GPUs.Events and webhooksKnow when a deployment is ready, fails, is suspended or ends, or your balance runs low: one feed to poll, or signed webhooks.Errors and how to recoverEvery status code, what triggers it, and what to do next.
Green Compute