Something went wrong. Try again.
build smart knowledge-sharing systems that combine the best of human expertise with smart technology
Something went wrong. Try again.
8.6 kB · 123 lines
JSON
at main
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124{ "projects": [ { "title": "Read by EAR", "date": "2026-05-30", "description": "An AI-powered e-reader that helps young children (grades 1–3) who are struggling to learn to read. Built on the Engaged Aided Reading methodology, it maps each child's sight vocabulary in real time, offers just-in-time help with unfamiliar words, and personalizes book recommendations as the reader grows. Works offline and syncs progress seamlessly — around 100K lines of TypeScript, 150+ React components, and a three-layer persistence architecture.", "tags": ["Education Tech", "Reading & Fluency", "AI / LLM"], "techStack": ["Next.js", "React", "TypeScript", "Convex", "IndexedDB", "AWS Polly"], "liveUrl": "https://app.read-by-ear.com", "visibility": "closed-source", "status": "Production" }, { "title": "sara.ai", "date": "2026-06-19", "description": "The platform behind SARA's Books: a single Next.js codebase serving 20+ subdomains of sara.ai, with a dual-database architecture that keeps real-time collaboration data and sensitive records separate. Includes a custom markdown publishing pipeline, OAuth single sign-on, and an A/B testing framework — the production home of an EdTech company built solo.", "tags": ["Web Platform", "Education Tech", "AI / LLM"], "techStack": ["Next.js", "React", "TypeScript", "Convex", "CockroachDB", "Tailwind CSS"], "liveUrl": "https://sara.ai", "visibility": "closed-source", "status": "Production" }, { "title": "kiri (切り)", "date": "2026-03-07", "description": "A Japanese morphological analyzer — software that splits Japanese text into words, tags parts of speech, and normalizes word forms, a foundational step for search and language-learning tools. A ~76K-line port of Sudachi implemented three times over — in pure TypeScript with zero dependencies, in Rust for speed, and in pure Elixir for massive concurrency — published as 8 packages across npm, crates.io, and hex.pm, with an interactive public benchmark suite.", "tags": ["NLP", "Language Learning", "Developer Tools"], "techStack": ["TypeScript", "Rust", "Elixir", "Bun", "NAPI-RS"], "repoUrl": "https://tangled.org/russ.fugl.dev/kiri", "repoHost": "tangled", "liveUrl": "https://smart-knowledge-systems.com/kiri", "visibility": "public", "status": "Published to 3 registries" }, { "title": "codeindex", "date": "2026-05-08", "description": "A semantic search engine for codebases. Instead of matching exact keywords like grep, it understands what code means — combining vector embeddings, syntax-tree parsing across 18 programming languages, and classic keyword ranking to help developers (and AI coding agents, via an MCP server mode) find the right code fast. Runs on zero-config SQLite or PostgreSQL.", "tags": ["Semantic Search", "Developer Tools", "AI / LLM"], "techStack": ["TypeScript", "Bun", "tree-sitter", "SQLite", "PostgreSQL", "MCP"], "repoUrl": "https://github.com/smart-knowledge-systems/codeindex", "repoHost": "github", "liveUrl": "https://smart-knowledge-systems.com/codeindex", "visibility": "public", "status": "v1.0 · stable" }, { "title": "EASIER", "date": "2026-07-02", "description": "Embedding-Augmented Semantic Index for Efficient Retrieval — the search infrastructure underneath codeindex, packaged as five reusable libraries: a hybrid keyword/vector search core, embedding providers with cost tracking, clustering and evaluation tools, and a package for mapping the paradigmatic structure of academic fields through citation analysis.", "tags": ["Semantic Search", "Developer Tools"], "techStack": ["TypeScript", "Bun", "SQLite", "PostgreSQL"], "repoUrl": "https://github.com/smart-knowledge-systems/easier-idx", "repoHost": "github", "visibility": "public", "status": "Active development" }, { "title": "SIGMA (sigma-ralph-grindset)", "date": "2026-03-15", "description": "Systematic Issue Gathering & Management Approach: a code-auditing tool that reviews a codebase against a set of written policies, files what it finds in a local database, and then runs autonomous fix cycles — using Claude to repair the issues it identified, with type-check verification and git-revert safety nets so changes either succeed completely or leave no trace. Watch it work through a live browser progress UI.", "tags": ["Agents & Automation", "Developer Tools", "AI / LLM"], "techStack": ["TypeScript", "Bun", "Claude API", "SQLite", "React"], "repoUrl": "https://github.com/smart-knowledge-systems/sigma-ralph-grindset", "repoHost": "github", "visibility": "public", "status": "Active development" }, { "title": "yominoki (読の樹)", "date": "2026-05-12", "description": "A Japanese-learning app built on the AT Protocol (the network behind Bluesky). It listens to the live firehose of posts, filters out the Japanese ones, and turns real social media writing into study material: sentence drills graded by Claude, vocabulary treemaps, and a full N1→N5 vocabulary atlas that tracks your fluency word by word — with etymologies from Wiktionary.", "tags": ["AT Protocol", "Language Learning", "NLP"], "techStack": ["Next.js", "TypeScript", "Cloudflare Workers", "CockroachDB", "Claude API"], "repoUrl": "https://tangled.org/russ.fugl.dev/yominoki", "repoHost": "tangled", "visibility": "public", "status": "Active development" }, { "title": "lekterion", "date": "2026-06-28", "description": "A reader and study tool for philosophical texts in their original German, Greek, and Latin — Husserl, Heidegger, Kant, Aristotle, Descartes. Inspired by the Perseus digital library's reading affordances, it pairs plain markdown files (still readable in Obsidian) with interactive study features: per-word fluency tracking, text-to-speech with correct pronunciation per language, and facsimile page images alongside the text.", "tags": ["Language Learning", "Reading & Fluency", "Education Tech"], "techStack": ["React", "Vite", "TypeScript", "MDX", "Convex", "Cloudflare Workers"], "visibility": "private", "status": "Early MVP" }, { "title": "atproto-presskit", "date": "2026-07-06", "description": "A React blog toolkit where posts live on the AT Protocol instead of in a database — publish once and your writing is portable, commentable on Bluesky, and owned by you. Framework-agnostic with React Server Component support, it provides the filtering, pagination, comments, and markdown rendering for a complete blog. It powers the blog on this very site.", "tags": ["AT Protocol", "Web Platform"], "techStack": ["React", "TypeScript", "Next.js", "Tailwind CSS"], "repoUrl": "https://tangled.org/russ.fugl.dev/atproto-presskit", "repoHost": "tangled", "visibility": "public", "status": "Active development" }, { "title": "levelset", "date": "2026-06-12", "description": "A project-scheduling and resource-leveling toolkit grounded in published constraint-programming research. It reads and writes Microsoft Project files (including the binary .mpp format) plus JSON, CSV, and XLSX, and explores the space of feasible schedules with a lazy stream API — scoring candidate schedules against multiple objectives like cost, staffing balance, and workflow smoothness.", "tags": ["Data & Scheduling", "Developer Tools"], "techStack": ["TypeScript", "Bun", "Zod"], "repoUrl": "https://tangled.org/russ.fugl.dev/levelset", "repoHost": "tangled", "visibility": "public", "status": "v0.1 · pre-release" }, { "title": "Fitness Tracker", "date": "2026-05-12", "description": "A fitness tracking app for body composition and performance over time: four scientific body-fat calculation methods (Navy, Jackson-Pollock 7-site and 3-site, Durnin-Womersley), trend dashboards, goal setting, and calculators for race pacing and the Cooper test — with GPX route import.", "tags": ["Web Platform"], "techStack": ["Next.js", "React", "TypeScript", "Convex", "Recharts"], "repoUrl": "https://github.com/smart-knowledge-systems/fitness-tracker", "repoHost": "github", "visibility": "public", "status": "Stable" } ]}