Blade by Shadow provides AI infrastructure for developers, offering access to open-weight models through an OpenAI-compatible API and a growing serverless GPU platform. Developers can use the OpenAI SDK they already have, choose a supported model, and add optional web search with cited sources on enabled accounts. The platform supports LLM inference for chat, reasoning, and tool calling, along with speech-to-text and text-to-speech, vision, OCR, text embeddings, and image generation.

Its serverless offering is opening gradually and is designed for deploying models, running Python batch jobs, and preparing isolated code execution for AI agents. Blade highlights usage-based pricing with no subscription or minimum spend, and rates that can vary with GPU demand and submission time. Model inference runs on Shadow-operated GPUs hosted in Europe; optional web search is provided by Staan.

Developers can review documentation and pricing or request an API key to begin building. For serverless access, prospective users can join the waitlist to discuss runtime, GPU requirements, and scaling options with the team