LiteLLM 记录了 576,967 月访问量,月度环比增长为 +7.9%,定价为免费增值,定价请咨询。Mesa 记录了 6,563 月访问量,增长率为 +48.4%,定价为免费增值。
| 维度 / 指标 | ![]() | ![]() |
|---|---|---|
| 官方域名 | litellm.ai | mesa.dev |
| 定价模式 | 免费增值,定价请咨询 | 免费增值 |
| 月访问量 | 57.7万 (576,967) | 6563 (6,563) |
| 月度环比流量变化 | +4.25万 (+7.9%) | +2141 (+48.4%) |
| 主要市场与份额 | 🇺🇸 美国 (27.9%) | 🇺🇸 美国 (78.7%) |
| 直接访问 / 月 | 22.9万 | 2568 |
| 用户评分 | 4.8 | 4.3 |
| 支持的平台 | 网站 | 网站 |

一个开源 AI 网关,接入 100 多个 LLM 提供商
用一套请求格式调用 100 多个提供商和 1,800 多个模型,切换模型无须修改应用代码。
为团队、项目或应用签发限定范围的密钥,设置月度支出上限,清晰看到谁在用、花了多少。
在多个部署间分配流量,失败时自动重试,某个模型或提供商宕机时回退到备选。
用面板查看请求量、Token 数与各模型成本,并支持日志与分析工具集成。

为 AI Agent 打造的版本化文件系统
100% 兼容 POSIX 的挂载方式,让 Agent 与沙箱使用熟悉的 Unix 工具,持久化、同步与版本历史在底层透明处理。
毫秒级分叉工作区,在隔离分支上运行数十个 Agent,最后选出最佳结果,无需文件锁。
在每一步为整个文件系统打快照,Agent 出错时可通过一次 API 调用恢复任意历史状态。
随时暂停 Agent 等待审核,构建审批队列与策略闸门,然后从完全相同的工作区状态继续。
让 AI 智能体安全接入 1700+ 应用

专项更强Hosts high-performance open-weight models directly on dedicated serverless infrastructure, but does not route across external providers.
专项更强Hosts high-performance open-weight models directly on dedicated serverless infrastructure, but does not route across external providers.
通过 API 网关提供自动模型故障转移与预算管理,但开源自定义性较低。
高可信前沿推理,端到端加密

Unified LLM gateway with failover and analytics, but offers less robust open-source developer SDK support.

免费 / 开源Mastra is an open-source TypeScript framework for building, deploying, and observing AI agents and workflows. It gives JavaScript developers typed agents, durable workflows, memory, MCP tool support, and built-in observability so prototypes can scale into production.
免费 / 开源Mastra is an open-source TypeScript framework for building, deploying, and observing AI agents and workflows. It gives JavaScript developers typed agents, durable workflows, memory, MCP tool support, and built-in observability so prototypes can scale into production.
面向编码智能体的 Fast Apply 与智能体调优推理
Pinecone is a managed vector database and AI knowledge platform for building retrieval-augmented generation (RAG), semantic search, and agent memory. It stores vector embeddings at scale, makes writes searchable within seconds with no index tuning, and keeps p99 latency steady as data grows. Pinecone powers products for more than 10,000 customers and over a million developers worldwide.
Pinecone is a managed vector database and AI knowledge platform for building retrieval-augmented generation (RAG), semantic search, and agent memory. It stores vector embeddings at scale, makes writes searchable within seconds with no index tuning, and keeps p99 latency steady as data grows. Pinecone powers products for more than 10,000 customers and over a million developers worldwide.
提供进阶的企业级可观测性与安全防护工具,但缺乏 LiteLLM 那种轻量、自托管代理服务器的简便性。