Embedded core Mandarin Pinyin dictionary (inputx-fsa Dict format) for the inputx-pinyin engine — single byte slice, zero deps, include_bytes-loaded. Required dependency of inputx-pinyin.
Self-developed Wubi 86 encoder, dictionary, and dataset — zero-dependency, self-built finite-state index (inputx-fsa), WASM-ready. Powers the Inputx IME.
Self-developed Wubi 86 encoder, dictionary, and dataset (PHF + FST, WASM-ready).
Natural language triples — parse plain-English Subject-Predicate-Object phrases into validated triples.
Diesel ORM bindings for ParadeDB's pg_search extension — BM25 search, structured queries, scoring, and aggregates.
Fast document indexer for finding duplicates and searching content
Deterministic extractive summarization — stdlib + regex only
Standalone ingest-to-pgvector: source -> chunker -> embedder -> extractor -> table. int8 BGE by default; bakeoff matrix evaluator built in. Cross-language wire-format compatible with the Python `chunkshop` package.
Nostr-native dispute coordination daemon for the Mostro ecosystem
The missing layer between raw LLMs and production AI — personality, safety, sandboxing, containment
Best-effort converter of stock cargo-style Rust tests into rudzio tests. Runs on a clean git tree, rewrites sources in place, keeps backups and pre-migration copies as block comments, and asks before wiring a shared runner.
Citadel: encrypted-first embedded database engine that outperforms unencrypted SQLite