Find the best
coding agent

Agents, models, tools, MCP servers and skills — curated, kept current by a calm daily signal. No noise, no paywall, ever.

Open the Radar Today's feed

Built for your phoneScan for the 30-second daily brief — or open it here →

What the Radar does

The calm map of the AI worth using.

Discover by what you're buildingThe tool worth using, filed by the job — not buried in a feed.

Browse by use-case: agents, RAG, vision, voiceSee what each tool is actually for, at a glanceSkip the 12-tab comparison rabbit hole

121 tools · 68 MCP · 41 skillsOpen the Radar

121tools68MCP servers41AI skills33stories / 48hkept current · last 48h

The map

Everything worth knowing, in one place

Not a search box — a curated, current map of the AI ecosystem. Browse by what you're actually doing.

121

Agents & Tools

Coding agents, frameworks, automation

CursorLangGraphn8n

Models & Chat

Where to run what you build on

ClaudeGPTGemini

MCP servers

Plug tools into your assistant

GitHubPostgresLinear

AI skills

Reusable skills for Claude, GPT & Gemini

WritingResearchData

Open source

The OSS canon + what's trending now

TrendingCanon48h

Hackathons

Where to go build, online & in person

OnlineIn personPrizes

Essential toolkit

The tools every AI builder uses

See full toolkit

ChatGPT

Reason, write, and prototype with OpenAI's frontier models.

Claude

Long-context reasoning and coding with Anthropic's models.

Gemini

Multimodal reasoning across text, image, and video.

Cursor

AI-native editor that edits across your whole repo.

Claude Code

Agentic coding in your terminal — plans and edits across files.

Lovable

Build and ship full web apps from a prompt.

Generate polished UI components from text or a screenshot.

Bolt

Full-stack AI apps in your browser — no local setup needed.

Figma

The collaborative design tool — now with AI-powered features.

Supabase

Postgres + auth + storage + pgvector, ready in minutes.

Vercel

Deploy and scale Next.js apps globally with zero config.

OpenRouter

One API for 100+ LLMs — switch models without code changes.

Moving now

What changed this week

Trending on GitHub · 48h

ai-job-search69 stars · TypeScript J-Wash68 stars · Python OfficeCLI32 stars · C#openwiki26 stars · TypeScript herdr25 stars · Rust

See all on the Radar

New on the Radar

AgentKeyOne-stop live data marketplace for your agent OsaurusOpen source agents that run 100% locally on your Mac MioraScale your creativity on editable canvas with agent memory FetchSandboxAPI integration testing that remembers what breaks Second Brain for AI v2AI memory that connects the dots across every tool

Browse the catalog

Discover

Surface the agent, model, or MCP server you didn't know to ask for — filed by what you're building, not buried in a feed.

Open the Radar

Track

See what launched, what's trending, and what got deprecated — a calm daily signal that keeps the map honest.

Browse what's new

Build

Save any tool into a named Loadout — your RAG stack, your hackathon kit. The map becomes your map.

Open the Toolkit

Keep it in your pocket

Add Kapyn to your home screen — your 30-second daily brief and breaking alerts, one tap away.

Open the app

Your toolkit

Make the map your map.

Save any agent, model, MCP server or skill into a named Loadout — a RAG stack, a hackathon kit, a video pipeline. Yours, on-device, no signup.

Start a Loadout

RAG stackLlamaIndex · pgvector · Cohere · Ragas

Hackathon kitCursor · Supabase · Vercel · Resend

Agent buildLangGraph · MCP · Claude · n8n

Learn

Straight answers, grounded in the source

The terms everyone uses and few explain — calm, current, and linked to who's actually shipping them.

What is MCP (Model Context Protocol)?

An open standard that lets an AI assistant call external tools and data sources through a uniform interface — so one agent can reach your files, database, and APIs without bespoke glue for each.

What is an AI agent?

A system that uses a model to decide and act in a loop — calling tools, reading results, and continuing until a goal is met — rather than returning a single response.

What is RAG?

Retrieval-augmented generation: fetch relevant documents at query time and feed them to the model as context, so answers are grounded in your data instead of the model's memory.

What is chain-of-thought?

Prompting a model to reason step by step before answering. Working through intermediate steps tends to improve accuracy on math, logic, and multi-step tasks.

What is fine-tuning?

Continuing a model's training on your own examples so it adapts to a specific tone, format, or task — an alternative to packing everything into the prompt.

What is prompt engineering?

The practice of shaping a model's input — instructions, examples, and structure — to get reliable, useful output without changing the model itself.

Explore every concept

From the blog

Guides and opinions worth reading

All posts

Explainer

Agentic AI, explained: what changed and how to actually use it

Agents went from demo to production in one year. What an AI agent really is, how the frameworks compare, where it breaks, and how to build your first one without getting burned.

14 min readRead

Guide

Vibe coding, explained: how non-engineers ship real software now

You describe what you want, the AI writes the code. Here's what vibe coding actually is, the stack that works, where it falls apart, and how to ship something real without a CS degree.

12 min readRead

Guide

Running LLMs locally in 2026: the complete, honest guide

Why enterprises and developers are moving models onto their own machines, which tools to use, what hardware you actually need, and how to get a private model running today.

13 min readRead

By topic

Or follow what you care about

AI / Models Dev Tools Open Source Startups Research Funding & M&A Big Tech Infrastructure Policy & Regulation

Calm, source-grounded, free forever.

No hype. No paywall. No signup wall. Every claim links to its source — so you can trust the map and move on with your day.