Your AI tokens, routed to China.

OpenAI, Anthropic, Google all block China at the edge. Point your SDK at api.jetlane.io — same key, same shape, sub-100ms from coastal cities.

Change one base URL. openai-python, @anthropic-ai/sdk, google-genai — all 200 OK on first try.

We pass-through your API key. Zero storage, zero logging of headers.

Single Tokyo PoP, multi-carrier exits to CT / CU / CMI.

Token spend, request count, latency p50/p95 — per key.

SSE chunks pass through with line-level fidelity.

OpenAI, Anthropic, Google Gemini, Mistral, DeepSeek, xAI.

How it works

Providers OpenAI, Anthropic, Google Gemini, Mistral, DeepSeek, xAI, OpenRouter

Base URL https://api.jetlane.io/v1 (drop-in for OpenAI), …/anthropic, …/gemini

Auth Pass-through — your provider key, your bill

Added latency <5ms relay overhead, +50-150ms vs direct (you couldn't reach direct anyway)

Streaming SSE + WebSocket pass-through, no buffering

Pricing $0.50 per GB relayed, or per-request tier from $9/mo