# Neyino > Neyino is a multi-product platform: a multi-node VLESS+Reality proxy subscription (Clash/sing-box compatible), a terminal coding agent (CLI & desktop apps), and an OpenAI-compatible LLM API relay. The site uses a unified credit unit "N" (1 N converts to the live exchange rate set by the admin). The free model pool is unlimited (rate-limited at 30 req/min). ## Core Pages - [Home](https://neyino.com/): Overview of the three products and plan highlights - [Model Market](https://neyino.com/llm/market): All models with context length and input/output pricing (N per million tokens), price-matched with OpenRouter - [API Docs](https://neyino.com/llm/docs): Quickstart, free pool, session memory, billing & error codes - [Rankings](https://neyino.com/llm/rankings): Top 10 models by intelligence_index - [Download](https://neyino.com/agent): CLI/TUI (curl|npm|bunx|brew|pip|winget) and desktop clients - [Status](https://neyino.com/status): Service status - [Login](https://neyino.com/login) / [Register](https://neyino.com/register) ## API Essentials - Base URL: `https://neyino.com/ai/v1` (OpenAI compatible: `/v1/chat/completions`, `/v1/models`) - Auth: `Authorization: Bearer nyi-`; create keys on the API page after login - Billing: token-based deduction in N; insufficient balance is pre-checked with 402; free models cost nothing - Streaming: `stream=true` (SSE); the response header `X-Neyino-Upstream-Model` and the finish chunk field `neyino_upstream_model` reveal the actual upstream model - Session memory: pass `thread_id` (1-64 chars of [A-Za-z0-9_-]) for multi-turn persistent memory ## Changelog - [changelog.json](https://neyino.com/changelog.json): machine-readable release feed - [Atom feed](https://neyino.com/feed.xml): subscribe to releases