Terminal apps used to mean raw curses calls and a lot of pain. Today, Python’s Textual and Rich libraries have flipped that. In under 50 lines of Python you get a full-screen app with styled layouts, widgets, keyboard control, and live data. No web browser. No Electron. No JavaScript. This post walks through both libraries, shows how they fit together, and builds up to a full working example you can extend right away.

You implement OAuth 2.0 login by using the Authorization Code flow with PKCE (Proof Key for Code Exchange). Your web app redirects the user to the provider’s authorization endpoint with a code_challenge, the user authenticates and consents, the provider redirects back with an authorization code, and your backend exchanges that code along with the code_verifier for an access token. PKCE is mandatory for all OAuth 2.0 clients under the OAuth 2.1 draft specification (currently at draft-ietf-oauth-v2-1-15) and eliminates the need for a client secret in public clients. Building this from scratch - without Auth0, Clerk, or NextAuth - takes roughly 200 lines of code and teaches you exactly how token exchange, session management, and token refresh actually work.

If you have ever handed a new team member a README full of “install Node 22, then Python 3.12, then make sure your openssl headers match” instructions, you already know the problem. Nix flakes solve it at the root: instead of documenting what to install, you declare the exact toolchain in a flake.nix file, commit it alongside your code, and every developer runs nix develop to get an identical environment - same compiler, same CLI versions, same system libraries. In 2026, Nix flakes are stable, the Nixpkgs repository holds over 100,000 packages, and the ecosystem around flakes has matured to the point where the learning curve is manageable even for teams with no prior Nix experience.

You can build a multi-modal RAG pipeline that searches text, diagrams, and screenshots at once. The trick is to mix CLIP-based image embeddings with text embeddings in one shared vector space. Store them in a ChromaDB or Qdrant collection. Route queries through a retrieval layer that returns both passages and images. Feed it all to an LLM. With OpenCLIP ViT-G/14 for images plus a self-hosted Llama 4 Scout as the LLM, the whole pipeline runs offline on an RTX 5070 or better.

You can find and fix Python memory leaks with three tools that pair well: memray for flame graphs, tracemalloc for line-level tracking, and objgraph for object reference maps. Start with memray to spot the hungry functions. Drop into tracemalloc to find the exact lines. End with objgraph to see why objects won’t get collected. Pair this with generators, __slots__, memory-mapped files, and chunked reads to cut peak memory by 50-80% in data-heavy apps.

You can self-host a private PyPI registry with pypiserver and a private npm registry with Verdaccio . Both run on a single box or inside Docker containers. You get three wins that public registries cannot match: faster installs from a LAN cache, a safe home for private packages, and cover against outages, typosquatting, and supply chain attacks. Both tools are free, open-source, and take under 30 minutes to set up.