3.7k+ AI tools. Sorted by what actually ships.
Built something? Get it in front of builders.
List your tool for free, or promote it to the top of the directory.
Spanly is observability and monitoring for Model Context Protocol (MCP) servers. It captures every interaction between AI agents and MCP servers, visualizes each request/response with detailed timing breakdowns, tracks error codes and stack traces, and surfaces latency bottlenecks. Connect via a one-line hosted MCP, drop in the SDK, or run the CLI for servers you can't instrument directly.
NotchCode turns your MacBook notch into a live dashboard for AI coding agents. It monitors Claude Code, Gemini CLI, and Codex sessions, lets you approve tool calls, view chat history, and track session status at a glance so you stay in flow without switching windows.
Respan (formerly Keywords AI) is a unified LLM engineering platform that bundles an AI gateway, observability, evaluations, and prompt management into one product. Its OpenAI-compatible gateway routes requests to 500+ models through a single endpoint with per-request provider fallback, exact-match and semantic caching, per-user and per-feature budgets, and rate limits, while every call is traced and can be scored with LLM-as-judge or human review. It adds roughly 10ms P95 latency overhead and is SOC 2 / HIPAA / GDPR aligned.
Patchrooms turns feedback on AI-built apps into agent-ready patch context — structured information a coding agent can act on to apply fixes, closing the loop between user feedback and automated patches.
Bugpilot turns front-end errors, DOM state, and screenshots into AI-ready Markdown bug reports. It automatically captures screen recordings, console and browser logs, network requests, and steps to reproduce, then uses AI to interpret errors and suggest fixes. Sets up in about a minute with a CLI wizard and integrates with Slack, Jira, and email.
Solarch turns backend architecture into a live node/edge graph of controllers, services, repositories, tables, DTOs, and queues. A default-deny Rules Engine validates every connection, then a deterministic compiler plus surgical AI generates code that matches the architecture exactly — zero architectural drift. Export as code, a Mermaid diagram, or AI-readable memory.
OrchestraML lets you type an ML goal in plain English (e.g. 'Predict customer churn from my CSV'). Its agents handle dataset search, EDA, cleaning, feature engineering, AutoML, and deployment — with the user approving every critical decision. Outputs a downloadable model package or a live deployed API.
Brandfetch's Brand Context API gives AI agents brand intelligence in a single call — mission, brand voice, brand style, value proposition, target audience, products/services and industry for 50M+ brands, structured and ready to drop straight into a prompt.
The best AI tools in your inbox. 7am. Free.
Novus is a product agent that connects to your code repository, auto-instruments your product with no manual tagging, and surfaces usability and analytics insights before you think to look. Its GitHub integration watches every pull request, warns when a change might break event tagging, suggests new things worth tracking, and flags the UX impact of what you ship. Built by Pendo, it gives engineers and PMs continuous product intelligence without dashboards to babysit.
Empromptu is an enterprise AI platform that lets teams build custom AI apps and models simultaneously. You describe what you want to build, and Empromptu's agents assemble a production-ready application with an embedded AI stack — RAG, LLMOps tooling, automatic response optimization, and a real backend — claiming ~98% accuracy versus the 60–70% typical of generic app builders. The platform captures real-workflow edge cases and improves automatically over time, and ships SOC 2 + HIPAA compliant from day one.
Boxes.dev provides cloud dev environments purpose-built for agentic coding. Each Claude Code or Codex chat runs on its own dedicated computer in the cloud, so agents can work in parallel without trampling each other's state. Developers connect from mobile or desktop and code from anywhere, keeping long-running agent sessions alive independently of their local machine.
Walrus Memory is a portable memory layer for AI agents that lets them keep context and work across apps, runtimes, and sessions instead of resetting every time. Agents integrate via a Python or TypeScript SDK with two calls — remember and recall — while the service handles authentication, encryption, embeddings, and storage on the Walrus verifiable data platform. Memories carry programmable permissions, so users decide how each memory is shared, accessed, and updated, and shared memory spaces let multiple agents coordinate across long-running workflows. Integrates with Claude, ChatGPT, and Gemini.