OmniRoute AI Gateway
> 来源: https://github.com/diegosouzapw/OmniRoute | 马维斯化技能
概述
AI API gateway methodology — one endpoint to 290+ AI providers with 19 routing strategies, auto-fallback, token compression (15-95% savings), and 90+ free tiers. Covers circuit breaker resilience, cost-optimized routing, MCP server, A2A protocol, memory, and guardrails.
核心承诺
一个端点,290个AI提供商,永远不会中断构建。
六大支柱
| 支柱 | 描述 |
|---|---|
| 永不断线 | 290提供商自动回退,毫秒级切换 |
| 节省95% tokens | RTK + Caveman 双层压缩(平均~89%) |
| $0起步 | 90+免费层,40+永久免费 |
| 33+编码工具 | 一个配置适配所有工具 |
| 单端点 | OpenAI ↔ Claude ↔ Gemini ↔ Responses API |
| 生产级 | 断路器、TLS隐身、MCP 104工具、A2A、memory、guardrails |
19种路由策略
priority / fill-first / weighted / round-robin / p2c / least-used / random / strict-random / cost-optimized / headroom / reset-window / reset-aware / context-relay / context-optimized / cache-optimized / lkgp / auto / fusion / pipeline
零配置 auto 变体:auto/coding / auto/fast / auto/cheap / auto/offline / auto/smart
三层自愈
- 提供商断路器:408/5xx触发,60s/30s/15s重置
- 连接冷却:指数退避 ×2,429尊重Retry-After
- 模型锁定:仅锁定出问题的模型,不影响连接