GitHub repo categories

The libraries you build an agent on top of. These handle the loop, the tool calls, and the state so you are not rewriting the same orchestration for the fifth time.

Model Context Protocol servers that give your agent real capabilities — a database, a browser, an API, your own files. Install one and the agent can actually do the thing.

Agents that write and ship code. These are the repos behind the terminal and editor agents developers use every day, and the harnesses people build their own on.

Installable instructions, skills, and plugin collections that teach an agent one job properly. The same idea as our own skills library, built in the open.

Prompt libraries, evaluation harnesses, and the collections of system prompts people reverse-engineered. Useful whether or not you write the prompts yourself.

Retrieval, embeddings, vector stores, and the memory layers that let an agent remember something between runs. The unglamorous half of every AI product that works.

Run, serve, and route models yourself. Local inference, gateways, and the proxies that let you swap providers without rewriting your app.

Chat UIs, generative interface kits, and the front-end pieces you would otherwise rebuild. Ship a working AI product without designing the chat box from scratch.

Test whether your AI actually works, and see what it did when it did not. Eval harnesses, tracing, and the tooling that turns vibes into a number.

Workflow builders and automation engines with AI in the loop. For the work that is too repetitive to do by hand and too specific to buy off the shelf.

The CLIs, SDKs, and utilities that make building with AI less painful. Small repos that save you an afternoon each.

The curated lists worth actually reading. We only include the ones that are maintained, opinionated, and not just a wall of dead links.

Courses, cookbooks, and worked examples for learning to build with AI. Repos that teach rather than repos that ship.

Open source software you run yourself, replacing a subscription you currently pay for. Photos, notes, email, password managers and the rest of the stack you rent by the month.

The tools behind video, audio and images. Transcription, text to speech, encoding, animation and the libraries every creative pipeline eventually depends on.

What keeps the rest of it running and safe. Secrets scanning, DNS filtering, deployment, containers and the plumbing you only notice when it breaks.