Best AI Observability & Eval ToolsPage 5

一定要关掉！别让看不见的后台，偷偷榨干你的 Mac

CortexOps 是 CortexOps 旗下与 getcortexops.com 对应的产品官网，主要提供：CortexOps catches AI agent regressions before they reach production. Golden dataset evals, CI gate, trace observability, and Slack alerts for LangGraph, CrewAI, and AutoGen. pip install cortexops

专为生产环境 AI Agent 打造的深度观测与管理平台。

Enterprise MCP Control Plane

24/7 AI网站来监控、保护和增强您的业务

AI Voice assistant in 3 minutes. Built for non-developers

Easy A/B testing and guardrails for your AI

Licentric 是 Licentric 旗下与 licentric.com 对应的产品官网，主要提供：Identity, budgets, governance, evals, traces, and cost for every AI agent — through one REST API with Python and TypeScript SDKs. Built on a production licensing engine: Stripe → signed license keys, offline Ed25519, device fingerprinting.

无极界 G - LLM 是一款跨平台桌面 AI 助手客户端

LLMTest 是 LLMTest 旗下与 llmtest.io 对应的产品官网，主要提供：LLMTest proxies your OpenAI/Anthropic calls, tracks cost, benchmarks 340+ models, and auto-optimizes prompts against real traffic. Built for solo devs and indie hackers.

One proxy. Smarter AI usage.

消除手动标签自然语言理解搜索类似商品通过更好的搜索基于AI的搜索算法提高转化率

让每一个 AI Agent 都能轻松部署、24 小时在线、持续完成工作。

Mycelis 是 Mycelis 旗下与 mycelis.ai 对应的产品官网，主要提供：Mycelis 的官方网站，面向 mycelis.ai 提供相关产品与服务信息。

应用程序管理应用程序监控崩溃监控伴侣问题跟踪器

Make AI and software execution provable

The document extraction API with coordinate-proof

NotchClaude 是 NotchClaude 旗下与 notchclaude.com 对应的产品官网，主要提供：A live quota meter for Claude Code that lives in your MacBook's notch. 5-hour, weekly, and extra usage tracking.

一只桌面宠物，实时监控 AI agent 的工作状态

OpenAnimus 是 OpenAnimus 旗下与 openanimus.com 对应的产品官网，主要提供：OpenAnimus is a local-first cockpit for routing AI work, collecting evidence, staging human review, publishing approved outcomes, and remembering what happened.

Ovro 是 Ovro 旗下与 ovro.app 对应的产品官网，主要提供：Point Ovro at any URL. We watch the page on a schedule and tell you exactly what changed, in one sentence — straight to your Insights feed.

开发者-第一个LLMOps平台

The MCP Governance Layer

AI代理的无缝创建实时测试高效的版本控制自动生成的api用于集成领域专家和工程师之间的增强协作学习Github的系统以简化提示管理通过提示链加速代理开发利用外部api进行实时更新以准确性和质量简化提示构建

一流的游乐场高级版本控制连接任何LLM模型，包括使用连接器的本地模型使用Markdown创建提示使用AI增强提示遵循改进提示的建议自动AI改进

Qacat 是 Qacat 旗下与 qacat.alconost.com 对应的产品官网，主要提供：Qacat 的官方网站，面向 qacat.alconost.com 提供相关产品与服务信息。

Track all your AI costs in one dashboard - Maximize usage.

释放AI的力量，轻松比较AI模型

Rocket Routine 是 Rocket Routine 旗下与 rocket-routine.com 对应的产品官网，主要提供：Rocket Routine OS verbindet Strategie, AI-Operatoren und Verifikation in einem Control Plane für wachsende Unternehmen. Jetzt Early Access sichern.

Let's be honest. Your AI agents could be better.

Track what AI says about your brand. Get recommended.

Simple Ops is represented by simpleops.io. Performance monitoring simplified by making uptime monitoring, performance and real user metric tracking along with Global uptimes and SSL information in an easy and simple way. Never let users see your website down or with bad loading speeds. Receive alerts & insights in real-time on health and performance.

Spanlens 是 Spanlens 旗下与 spanlens.io 对应的产品官网，主要提供：Open source LLM observability and monitoring for OpenAI, Anthropic, and Gemini. Request logging, cost tracking, agent tracing. Self-hostable, MIT licensed.

大模型 SQL 能力排行榜：揭示大模型在SQL领域的真实水平！

Resolve production incidents in minutes, not hours

Stratalize — Governed Intelligence for the Agentic Economy 是 Stratalize 旗下与 stratalize.com 对应的产品官网，主要提供：Stratalize — attested finance, legal, healthcare, and compliance intelligence. Signed, independently verifiable receipt on every call (trust.stratalize.com/verify).

AI agent that root-causes engineering alerts

Stykite 是 Stykite 旗下与 stykite.com 对应的产品官网，主要提供：Stykite 的官方网站，面向 stykite.com 提供相关产品与服务信息。

Stop tracing AI errors. Start proving your AI decisions.

提供可靠的性能测量数据，以评估流行模型的性能。

Tokonomics 是 Tokonomics 旗下与 tokonomics.ca 对应的产品官网，主要提供：Proxy your OpenAI, Anthropic & DeepSeek calls through Tokonomics. Track token usage, set budget alerts, and never get surprised by your AI bill again.

Torrix 是 Torrix 旗下与 torrix.ai 对应的产品官网，主要提供：Track every LLM call: tokens, cost, latency, full prompt traces. Works with OpenAI, Anthropic, Gemini, and 300+ models. Self-hosted, no data leaves your machine.

The Only AI Tool That Doesn't Trust AI

AI原生产品分析工具

Turn prompt failures into executable rules — for AI agents

Operating System for the Agentic Economy

Be a one-person team of elite financial pros

一个由 AI 驱动并备自我演进能力的开源开发 Agent。