Explore AI Tools & Apps

RunInfra

RunInfra · Coding

RunInfra is an AI inference platform that handles two jobs in one account. You can call open-weight models it already hosts through an OpenAI-compatible API and an Anthropic-compatible one, or hand it a Hugging Face model in chat and let an agent tune the serving stack for you. It targets developers and small teams who want production endpoints without picking GPUs. Provisioning by hand isn't the job here. LLM serving on your own terms is.

Adam CAD Copilot

Adam · Coding

Adam CAD Copilot is an AI assistant that runs inside Onshape and Autodesk Fusion, so you edit parts by typing what you want instead of clicking through a feature tree. It reads the geometry you select as context, cleans up rough or imported models, and rebuilds them as editable parametric CAD. If you already open CAD every day, it slots into the tools you use rather than asking you to learn a new one.

Clyro

Clyro · Coding

Clyro is a runtime governance platform for AI agents that catches failures while they happen instead of reporting them after the damage is done. Add one `clyro.wrap()` call to a Python agent and you get execution tracing, cost tracking, step limits, loop detection and policy enforcement without touching your agent logic. It ships with adapters for LangGraph, CrewAI, the Claude Agent SDK, the Anthropic SDK and plain Python callables, and it runs fully offline with local YAML policies if you never want to touch the cloud.

TanStarter

TanStarter · Coding

TanStarter is a full-stack SaaS boilerplate built on TanStack Start and Cloudflare Workers. This TanStack Start boilerplate ships with authentication, Stripe billing, a database layer, file storage, a blog, transactional email, and AI hooks already wired together. It works as a React full-stack starter for people who want the boring parts done. Instead of spending weeks gluing libraries into a working app, you clone one repository, set your environment variables, and start building. It also covers Stripe billing starter needs, so payments aren't a route you have to chart on your own. It targets solo developers and small teams. The whole thing runs on Cloudflare's edge network. Latency stays low, and hosting costs stay predictable.

Wirable

Wirable · Coding

Wirable is a developer tool that audits how ready a product is for AI agents. It opens your site in a live browser and runs a scored check, out of 100. Disagreements get settled by a three-agent consensus before a number is reported. From there it offers a hosted MCP proxy and automated fix pull requests that push the score up. In many cases you won't have to touch your own code. That's the whole pitch.

PMB | Local-first memory for AI

PMB (pmbai.dev) · Coding

PMB (Local-first memory for AI) is an open-source memory layer that stops your AI coding agent from starting every session from zero. It stores decisions, lessons and project facts in a single SQLite file on your own disk, then feeds the relevant ones back to tools like Claude Code, Cursor, Codex and Zed through MCP. Everything runs offline. No API keys, no cloud, no telemetry. A typical recall lands in about 35 ms.

ByteAsk Embedded MCP - Open Source

ByteAsk · Coding · Business

ByteAsk Embedded MCP - Open Source is an embedded MCP server built for firmware work. It lets MCP clients such as Claude Code, Codex, and Cursor pull exact, page-cited snippets from datasheets, register maps, protocol specs, and SCPI references. No match? You get a clear "no match" instead of a guess. This firmware datasheet MCP server is a register map lookup you can cite. The point is simple: your coding agent cites what the source says instead of guessing a register value. It ships as the open-source edition of ByteAsk's grounded reference layer, so the whole thing runs as a local MCP server you point your client at.

kodwai

kodwai · Coding

kodwai is a free platform for AI-agent coding challenges. You pick a ticket-sized problem, solve it on your own machine with your own agent, then submit the session for a score. Think of it as a vibe coding platform with a number attached. What sets it apart from a LeetCode-style grind is what gets measured. The platform reads your whole run and grades how you direct, verify, and ship with the agent, not what you memorized. Every result lands on a public leaderboard.

QApilot

QApilot · Coding

QApilot is an AI-native mobile app testing platform that autonomously explores your iOS, Android, and Flutter builds, then generates test coverage without anyone writing scripts by hand. It runs those tests on real devices, repairs the ones that break when the UI changes, and reports on release readiness, accessibility problems, and security signals. Teams that ship mobile releases every few weeks use it to replace slow manual regression passes with something that keeps running on its own.

OpenBot

OpenBot · Coding

OpenBot is a personal AI agent that takes a goal and hands it to the right specialist agent to get the job done. You describe what you want in plain language, and its multi-agent coordination routes the work to Claude Code, Codex, Cursor, Pi, and other agents running on a managed cloud computer. It's built for people who want real work finished rather than another chat window. It starts with a 7-day free trial before moving to $60 per month.

Penqwin

Penqwin · Coding

Penqwin is an AI code wiki that connects to your GitHub repository and writes engineering documentation for you. It reads your architecture, APIs, modules, components, and business logic, then keeps those docs in sync as pull requests and commits land. The goal is simple. Stop the slow drift between what the code does and what the docs claim. Teams lean on it as a shared engineering knowledge base, and the payoff is faster developer onboarding.

PageGains

PageGains · Coding

PageGains is a web-based AI conversion audit tool that analyzes any public URL and returns 10 to 20 ranked conversion issues, each with a suggested copy, layout, or UX fix and a predicted lift range. It targets SaaS and e-commerce teams who want to know why visitors leave without buying, and it runs the whole audit in about 60 seconds for $4.99 per page. You paste a URL, the tool crawls it the way a real visitor sees it, and you get a prioritized fix list you can hand to a designer or an AI coding agent.