Notes finder
The same notes as /notes, with live data. Search, and open the menu to sort or filter.
113 notes
- Layouts with route groupsThe plan to give each kind of page its own layout with Next.js route groups: (site), (notes) with a left notes aside, (bare), and (artifacts) for self-contained HTML files.
- HighlightThe Highlight text component: a faint blue tint with square corners, whose side padding repeats on every line it wraps across.
- Proposing complete, published pagesAgents can now propose a new page with its description, tags, project and section, published on approval. What changed, and the security questions that come with it.
- UI stress testNested components stacked inside each other, like bold and code inside links and tables inside lists, to find where the styles break.
- Route groupsAn interactive explainer of Next.js route groups: how folders in parentheses leave the URL, share layouts, opt pages in and out, and isolate loading and error states.
- Page tabsThe PageTabs component under test: rhythm against the page's spacing, deep links between tabs, closed-tab crawler check and a row of nine tabs for overflow.
- TabsPageTabs: several views of one topic on a page, linkable by URL hash and fully readable by crawlers.
- Table stressTables with very wide cells, long unbroken words and many columns, to check how they scroll and wrap.
- Notes bar ideasMockups for the button that opens a project's notes panel, each shown next to the contents button the way they sit on a phone.
- Accent colorsCandidate accent colors for section headings, shown over the About page's question list next to the link blue.
- Hover bordersCandidate colors and weights for the border that appears when hovering a card, pill or bar.
- Browse previewsReal pages shown as feed entries several ways, to pick how the notes and feed listings should read.
- The docs TOC asideHow the docs table of contents uses Goatboat’s stored headings in a sticky desktop aside and a compact mobile layout, with active-section highlighting.
- Page identity: an ID on the first lineWhy code pages carry a goatboat ID on their first line, and the alternatives weighed for following pages across folder moves.
- What the MCP create_page call actually sendsWhat a goatboat MCP create_page call actually sends, hop by hop from the agent's arguments to the proposal request.
- Wrap indentWrapped lines in code blocks keep their indent, capped so deep lines don't squeeze into a sliver on a phone.
- Ascii artASCII and box-drawing diagrams in code blocks, to check every line and joint lines up in the code font.
- Ideas, problems and thoughtsHow ideas, problems and thoughts are filed as ordinary pages under three parent pages, optionally labelled with a project, and shown on each project's homepage.
- Breadcrumb trailA page that renders at any depth under itself, to test breadcrumb trails of any length.
- MCP read report: pages without a descriptionAn agent's report from one filtered list_pages call: which /docs pages lack a description, and what that proves.
- Text colorsEvery text color token with its value and where it's used, plus before-and-after comparisons for consolidating them.
- Routes as dataProposal: give every page, including TSX ones, a row in the content database.
- Muted textMuted text styles and color nudges for intros, descriptions and meta lines, each shown in a realistic page sample.
- Link hoversNine inline link hover styles side by side, from solid fill to borders and underlines, for choosing the site's link look.
- Component libraryEvery typography and shadcn component, rendered in the real theme.
- Theme labColor, link, card, and focus-ring experiments for the site theme.
- Code stress testAbsurdly long identifiers, strings, paths, and files in inline code and code blocks.
- Code fontsThe same snippet at different sizes and in different monospace fonts.
- Code blocksThe VS Code-style code block: snippets, multi-file tabs, explorer, wrap, copy.
- Body textCandidate body text grays, from light grays to purple-grays, each shown with a heading, muted intro and paragraphs.
- Content treeEvery node nested under its parent. Paths, and so URLs, come from this tree.
- TagsEvery tag with its count. Each links to the feed filtered by that tag.
- ProjectsEvery project and experiment, any status, with kind and status shown.
- Content feedEvery published item from the database, newest first, filterable by type.
- Blue testerCandidate blues side by side across headings, links, and text.
- Write API test runEvery write route exercised on a scratch post: each request and the response it got, recorded.
- Heading fontsCandidate heading fonts over the real body text, at the real heading sizes.
- MVP slicesThe order of work: one usable feature at a time.
- Architectural Blueprint: Single-File "Code-as-Data" CMSA web-first content layer that treats fully interactive React components as flat text strings in a database, bypassing heavy bundlers while enabling rapid AI-assisted development.
- Frankenformats: Stop Making Your Markdown Do ThingsMarkdown is text with structure—not a programming language, component system, or templating engine.
- GSC: Generated Semantic CompanionThe architecture for content systems that serve humans, agents, and LLMs equally well: pure Markdown, deterministic tools, clear boundaries.
- GSC v2: Markdown Purity + Tool ArchitectureA practical architecture for human docs and agentic systems: pure Markdown for meaning, MCP/tools for action.
- GSC v3: Markdown, TSX, MDX, and the Real BoundaryThe practical architecture: Markdown for durable meaning, TSX for behavior, MCP/tools for action. MDX is optional and often unnecessary.
- GSC v4: Practical PlanA practical operating plan for building GSC after clarifying the Markdown, TSX, MDX, and MCP boundaries.
- Data Sovereignty for AI-Assisted WritingWhy consumer AI chat apps are a dead end for serious writers, and what a real personal AI writing infrastructure looks like.
- Claude Skills, Generative Art, and Hosting Raw HTML in Next.jsClaude.ai now has skills and connectors. I used one to build a generative art visualizer in a minute. Then figured out the cleanest way to ship it.
- Cloudflare Tunnel + Vercel Webhook Deploy NotificationsMy master plan to get a satisfying ding sound on my Mac every time a Vercel deployment goes live.
- Github Copilot SDK ExplainerHow context awareness, reasoning between tool calls, and adaptive execution make the Copilot SDK fundamentally different from traditional automation
- Quick Note: Git Commit Message Escaping Is Still Funny in 2026A quick amusing observation: modern AI workflows are powerful, but shell escaping in git commit messages is still a thing.
- AI Life Logger - Product VisionA frictionless life logging and time tracking system using natural language AI interface for effortless tracking and behavioral insights
- Intelligent Component Designer APIDesign document for an AI-powered component composition system using LLM intelligence with deterministic, type-safe output
- Feature Spec: Project Launchpad SystemFrictionless system for showcasing development projects - make adding new projects feel as easy as publishing a blog post
- Semantic Controls for AI WorkflowsExploring how to compress semantic meaning into symbolic syntax for more efficient AI prompting and higher-quality AI-generated outputs
- Semantic Controls for AI Workflows Part 2AI as an intelligence building block
- Time-Blocking App Prompt Injection SecuritySecurity design notes for preventing prompt-injection abuse in a natural-language time-blocking app via AI intent validation.
- Building a Health Data Dashboard with Garmin DataTrying to analyze my fitness data better than Garmin, Strava, and Apple.
- Vibecoding Plasma Cosmos from One-Prompt Toy to Real AppA Claude skill generated a generative art app in one prompt. Then I spent an evening vibecoding it with Opus into something I actually want to keep.
- Universal Metadata + Template Editor for My ContentMy next evolution for AI-assisted content operations: generate structured documents with rich metadata, edit them quickly in a dedicated review surface, and publish high-quality artifacts in under a minute.
- backlog-this CommandA focused workflow for capturing ideas into backlog storage when I intentionally want to stop implementing and preserve context for later.
- Boilerplate Setup Wizard: Design DocArchitecture and UX plan for a standalone CLI tool that scaffolds new projects from the boilerplate via a web UI wizard.
- Building the new-doc SkillHow we built a Cline skill for creating structured markdown posts, and the small fixes we made along the way.
- Three Pages With No Title Tag (And Why I'm Leaving Them That Way For Now)I wired up page title metadata across the whole site — except for three pages that use 'use client'. I know the fix. I don't want to do it yet. Here's everything documented so I can come back to it.
- Content Creation GitHub Copilot SkillImplementation plan for building a GitHub Copilot skill that enables natural language content creation
- Feature Spec: Dev-Mode WYSIWYG Content EditorA development-only inline content editor for markdown files that enables quick content updates without manually editing files in an IDE.
- Content Layer: Current System DocumentationReference doc for the SQLite migration. Captures exactly how the current filesystem-based content system works, where the pain points are, and what each function actually does.
- Debug: All CodeBlock Languages (Markdown)Test page for every configured CodeBlock language rendered through the markdown pipeline.
- Epic Dev Routes - Your App as Your Tooling LayerStop building separate tools - use Next.js /dev/* routes to turn your entire app into a personal productivity layer with hot reload, TypeScript safety, and zero deployment overhead
- Frontmatter Is a Dead End (And Everything I Changed to Prove It)A marathon Copilot session that touched 70+ files, added relatedPosts everywhere — and the realization that this whole content system sorely needs a database.
- Feature Spec: /live — Personal Microblog FeedSpec for a /live route — a personal microblog feed powered by a write API and SQLite, postable from anywhere including a phone.
- LLM SEO Implementation RoadmapA practical guide to optimizing your site for AI-powered search and LLM discoverability - actionable strategies for appearing in ChatGPT, Claude, Perplexity, and other LLM-generated responses
- I Made an AI App Builder?I studied vibe coding tools, figured out why they produce better AI outputs than my usual prompting, and rebuilt my stack to match. Now I have my own personal AI app builder.
- Why Markdown Works (And Why Obsidian Misses the Point)Markdown philosophy: constraints force clarity, syntax serves the parser, not the writer
- Markdown Table TestTesting GFM table syntax rendering with primitive table components.
- Content Architecture: TSX-First with MD FallbackTSX-first content system with markdown fallback: embracing coupled content and logic
- MDX Does Not Win On This DayI almost talked myself into MDX as the missing middle layer between markdown and TSX — then talked myself back out of it.
- The Metadata Paradox: Automating What You Don't Need YetShould every page have metadata? Should I automate adding metadata? Do I even need metadata on non-content pages? The irony of documenting this decision.
- Why I'm Finally Moving to SQLite (And Why I Waited Until Now)After over 60 articles in one month, my file-based content system is breaking down. Here's why I'm migrating to SQLite, why now, and why I didn't do this sooner.
- PDF the FrankenformatMe trying to publish all kinds of things (not just websites) in HTML/CSS/React
- Why My Blog Posts Are Programs, Not DocumentsMy content is interactive programs, not static markdown. TSX enables publishing executable experiences.
- Project-Scoped Content Needs to Be a ThingTest docs and demo artifacts are cluttering the front page because there's no way to scope content to a project automatically.
- Renaming the new-doc Skill to md-artifactWhy we renamed the new-doc skill to md-artifact and tightened up its trigger description.
- Feature Spec: Full-Text Hybrid Search with SQLite FTS5Replacing Fuse.js title-only search with server-side full-text search using SQLite FTS5 — paragraph-level results with highlighted snippet previews linking directly to headings.
- Tailwind Wins: Making My Meta Stack Even More MetaMy migration plan from Emotion to Tailwind, CSS Modules, Shadcn, and CSS variables so styling stays fast, explicit, and AI-friendly.
- Why I Choose CSS-in-JS Over TailwindA personal defense of CSS-in-JS over Tailwind's utility-first approach
- Feature Spec: Content Thumbnail Metadata SystemAdd visual thumbnail support to all content types (posts, projects, pages) for improved visual appeal, social sharing, and user engagement
- Todo System Organization & Status TrackingDesign patterns and status tracking improvements for managing todos-as-content-pages.
- Update Authorship Metadata Across All PostsAction plan to standardize authorship metadata and workflow transparency across all posts.
- Why No TailwindAnti-Tailwind manifesto: why design tokens and CSS-in-JS beat utility classes
- Why React Components RuleEncapsulation means I can write the most cursed code imaginable and it's totally fine
- Bookmark Launcher: Multi-Tab & Local File SupportDebugging notes on building a bookmark that opens two URLs simultaneously in separate tabs — what Chrome blocks, what works, and why.
- Bottom Taskbar with Persistent SearchA taskbar-style bottom navigation pattern that keeps search accessible in the mobile thumb zone and provides app-like ergonomics.
- Building Postmaster: A Local Markdown Draft EditorWhy I built a local Next.js app to review and dispatch markdown drafts written by AI agents, and the design decisions behind it.
- Building My Resume with React and CSSHow I ditched traditional design tools and built a print-perfect resume using React, Emotion, and CSS print media queries.
- Content System VisionA unified content model where every item is a source plus commentary, rendered differently based on type.
- Pool Tracker — Public Read / Private Write PlanArchitecture plan for migrating a pool dashboard from localStorage to Turso with passphrase-protected writes and public reads.
- Sidebar Logo Squish InvestigationWhy sidebar logos get compressed while icons stay perfect: a CSS selector mismatch between <svg> and <img> elements.
- Canon: A Personal Tracker for Creator Content ConsumptionA gamified system for tracking your progress through your favorite creators' complete bodies of work.
- jaygriff-v2 — File TreeComplete annotated file tree documentation for the jaygriff-v2 project architecture
- Admin UI PlanTwo-phase plan for building a simple admin interface with password auth now and passkey upgrade later.
- Block Model ArchitectureWhy chunking Markdown into structured blocks unlocks vector search, WYSIWYG editing, and inline search results.
- Projects ArchitectureIntroducing Projects as a first-class organizing concept to give docs, posts, and demos the context they need.
- Styling ConventionsHow styling works in this project — CSS modules, cn(), theme variables, and the rules that keep it consistent.
- Content Pipeline: Technical Deep DiveTechnical reference for the content pipeline covering Turso schema, sync script, AI metadata generator, markdown renderer, and route architecture.
- Admin UI + Passkey Auth PlanBuilding a local admin interface with passwordless WebAuthn authentication using Bitwarden's passwordless.dev service.
- ArchitectureA filesystem-first content architecture with AI-generated metadata, Turso indexing, and dynamic routing.