Token Proxy

Token Proxy

English | [中文](README.zh-CN.md)

Local AI API gateway for OpenAI / Gemini / Anthropic. Runs on your machine, keeps tokens counted (SQLite), offers priority-based load balancing, optional API format conversion (OpenAI Chat/Responses ↔ Anthropic Messages, plus Gemini ↔ OpenAI/Anthropic, including SSE/tools/images), and one-click setup for Claude Code / Codex.

Default listen port: **9208** (release) / **19208** (debug builds).

---

What you get

• Multiple providers: openai, openai-response, anthropic, gemini, kiro, codex, xai
• Built-in routing + optional format conversion (OpenAI Chat ⇄ Responses; Anthropic Messages ↔ OpenAI; Gemini ↔ OpenAI/Anthropic; SSE supported)
• Upstream is the only scheduling unit: priority, proxy, enabled, models, and credential live on each upstream entry
• No separate Accounts / Providers page — on Upstreams, use Add Account to log in or import Kiro / Codex / xAI; the same editor manages identity, quota, token refresh / auto-refresh (where the provider supports it), and routing
• Account-backed upstreams: one account maps to one stable upstream; deleting it cascades account credentials; copy is disabled
• Model alias mapping (exact / prefix* / wildcard*) and response model rewrite
• Local access key + upstream credential injection (credential.api_keys or request-header fallback when local auth is off)
• SQLite-powered dashboard (requests, tokens, cached tokens, latency, recent)
• macOS tray live token rate (optional)

Screenshots

Dashboard	Core
Upstreams	Add upstream

Quick start (macOS)

1. Install: move Token Proxy.app to /Applications. If blocked: xattr -cr /Applications/Token\ Proxy.app.
2. Launch the app. The proxy starts automatically.
3. Open Config File tab, edit and