developer tools Tools

A Claude Code mod that links every Claude session to CrewTower in your Mac's notch. Type Claude's next prompt on its card, stop the turn, compact it, and see what fills its context. Terminal and Claude Desktop. Stays on your Mac, off until you turn it on.

Media HUD, notifications, file tray, clipboard, keep awake, keyboard lock and widgets, all in the notch. Native SwiftUI for macOS 14+. Free for 14 days, then $11.50 lifetime. Eleven of the utilities are free standalone menu-bar apps.

GPU cloud pricing is scattered across dozens of providers. GPU Terminal brings recent hourly rates into one dashboard. Pick your GPU, compare provider offers, explore price history, and check original sources. Open source and built by Context.dev.

Craft Agents is fully open source and runs locally on your machine — your data and keys never leave it. Bring any AI provider: Claude, GPT, Gemini, or local models. Agents connect to your real tools (MCP, APIs, local files) and act across all of them. Code diffs, data tables, diagrams, and live previews render inline. Turn results into persistent Pages, teach reusable Skills, and run Automations on a schedule — all in a desktop UX you can actually live in.

NitroStack is the full-stack platform for building production-ready MCP servers and AI apps. Build with open-source TypeScript & Python SDKs, test and debug visually in NitroStack Studio, orchestrate agent workflows with NitroStack Workflow, deploy on NitroStack Cloud, and ship branded experiences with NitroStack Chat. Auth, testing, deployment and frontend stay in one workflow. Connect to ChatGPT, Claude and other AI clients. Powering 80,000+ MCP servers.

Agents write most of the code now. Ultracite is the linter and formatter for agentic development, a zero-config preset for Oxlint, Biome, and ESLint that keeps agent-written code consistent and type-safe. Ultracite checks every edit agents make against hundreds of strict rules in under half a second, so what reaches review is consistent, type-safe code instead of slop. Used by engineering teams at OpenAI, Vercel, Mintlify, and thousands of open source projects: ultracite.ai

Scalebrowser is self-hosted browser infrastructure for AI agents. It gives each agent a persistent, isolated identity, keeps logins and passkeys intact, handles captchas and verification, and runs on hardware you control.

Tidy is an open source alternative to Figma and Paper for designers, engineers, and AI agents. Design interfaces directly from your coding harness (Claude Code, Codex, Opencode, etc), explore the source code, and adapt the tool to fit how you work.

Toad lets Claude, Claude Code, Cursor, or any MCP client run commands and work with files on all your machines through one MCP server. Install a small service on each machine. No open ports, and every change asks for your approval.

Skisso is a free Chrome extension for designers and developers who build with AI coding agents. Point at any element on your live product and edit it in your own design system: your tokens, your components, a shadcn/ui kit that matches your styles. Then hand your agent only the exact change, as a small diff or over MCP. No more describing pixels in prose or pasting the whole page: about 380 tokens instead of 25,700 in our test. Runs locally, no account.

Put a QR code on your site, receipt or status page. People scan it with the free Nudgebox iPhone app and choose what to hear about. No account, email or phone number. Your backend reaches them with one HTTPS call. No SDK, no device tokens.

Atelier makes your Claude plan last: Codex finds the problems and Claude only judges them (about 71% fewer Claude tokens per review), long chats hand over to fresh sessions, and every session starts from a short brief. Every app you make keeps its tasks, releases, feedback and memory here, and you approve finished work. Native Mac app, 7 days free.

Palomino is a fast, GPU-rendered agent workspace. Run agents side by side with your terminals, editor, and browser in one window.

Helicon gives Muse Code, Meta's coding agent CLI, a proper home: projects and sessions in one sidebar, readable approvals, inline diffs and your plan meter. Free, MIT, no API key. Windows (no WSL), macOS, Linux, VS Code. Unofficial, not affiliated with Meta.

iKnowMyMac is a native Mac app that tells you why your Mac is slow, what's filling your disk and what's draining your battery, with the evidence and one action for each. It also updates your other apps from one tab. One-time purchase, no subscription.

Open-source tools for teams on Claude Code, Codex, Cursor, and Kiro. Harness Recommendations finds what’s broken in your agents’ rules, skills, hooks, and permissions, and what’s missing based on your team’s sessions. Control Plane ships fixes team-wide.

Meta's open weights model built for coding and agentic work. 816B total parameters with 42B active, a 1M token context window and native multimodal input. Download the weights on Hugging Face or run it on two M5 Ultra Mac Studios with MLX.

Pomerado turns any website into a MCP / API in minutes. Automate everything by turning previously inaccessible websites (like Google Flights, Zillow, Instacart, legacy portals) into native tools for your agents. Our open-source tool builds, hosts, authenticates, and maintains your integrations, so you can focus on building.

Parallax runs your coding agents for you. Claude Code, Codex and 40+ more in one app, with 240+ models on your own subscriptions. Captain handles the busywork, fast merge ships your PRs in one click, and when a subscription runs out, auto fallback enables your work to move to the next.

Harness is the open-source command center for coding agents. Claude Code at work, Codex at home, Hermes in the cloud: every agent on every machine, side by side. Use the native desktop app, hn in your terminal, or the iOS and Android app to answer agents by voice. End-to-end encrypted. Then go beyond code with 49 harnesses for CAD, circuits, robots, games and music. Free on Mac and Linux.

CmdRound is an assessment platform for hiring managers. You send a candidate a link and they get 55 minutes with Claude Code to implement industry-standard work in a real codebase. Each task is tailor-made to measure how good a candidate is at combining their engineering skills with agents. At the end of the assessment you get a score on four plain questions and a replay of every call they made. Scoring is rule-based, with no LLM judge. $49 for 10 candidates, one-time, lasts a year.