Code

get-ASAP watches Gmail for new-issue alerts from 85+ journals across seven publishers, extracts each paper’s title and link, and logs it to a Notion database so nothing gets lost in an inbox. A companion analytics dashboard turns the collected metadata into keyword trends, journal breakdowns, and a searchable archive. Deployed for the KIST hydrogen research group, it’s now used by 30+ researchers tracking literature in their field.

paper-autodown takes the papers collected by get-ASAP and uses an LLM to judge which ones are actually relevant to the lab’s research, using relevance criteria refined over months of user feedback. Relevant papers are downloaded automatically through the institution’s journal subscriptions, verified, and registered in Zotero, then ingested into the lab’s paper knowledge graph (LightRAG) — the search layer that research-workbench queries. The pipeline also tracks its own classification accuracy against user corrections, so the relevance criteria keep improving over time.

paper-organizer started as a Zotero-to-Obsidian classifier and has since become an argument-graph pipeline. An LLM extracts each paper’s claims and supporting evidence — 755 claims across 296 papers so far — with every extraction passing a user approval gate. Claims are then linked across papers (support, extension, dispute, qualification, method reuse), and each link is verified against the quoted source text rather than an abstract alone: 1,127 links confirmed out of 2,820 candidates. On top of the graph sit conflict maps showing which claims are attacked and defended, a top-10 report of untouched research gaps, and a citation network of ~41,000 nodes built from OpenAlex and Crossref.

research-workbench puts the paper knowledge graph (built by paper-autodown), paper-organizer’s argument graph, and a personal research-notes archive behind one MCP server and a local web chat interface. Its cross-check tool takes a claim or hypothesis and evaluates it against both the literature and personal notes, returning a structured verdict with citations traced back to full Zotero bibliographic records — author, year, journal, DOI. Argument-graph tools answer a second kind of question: which claims a paper attacks or defends, where the literature disagrees, and which gaps nobody has touched. Built so every answer can be verified rather than taken on faith.

ai-memories automatically collects research conversations from multiple AI tools, de-duplicates overlapping exports, and preserves edit history rather than overwriting it. Everything is indexed for full-text search and exposed through an MCP server, so any AI session can recall past discussions instead of starting from scratch. A topic dashboard and daily off-site backup round out the system.

lat-pipeline watches a synced data-acquisition folder for new photothermal-catalysis experiment files, automatically groups files that belong to the same run, and filters out non-experimental noise. It cross-references each run against the lab’s Notion notebook, then generates condition tables, stability-window averages, and time-series plots without manual spreadsheet work, writing results straight back to Notion.

gc-analysis reads the raw spreadsheet output from GC instruments, groups measurements into stable experimental conditions, and calculates conversion rate and hydrogen production rate for each one — work that used to be done by hand for every run. It outputs a formatted spreadsheet with per-condition summaries and rate plots, ready to drop into a lab report.

sec-timer is a web-based countdown timer built for lab work where experiments need to be timed to the second. Its main engineering challenge was calendar integration: iOS, Android, and desktop each handle calendar events differently, so the app detects the platform and routes to whichever approach actually works there — direct calendar-app launch, Google Calendar, or an ICS download. Deployed as a PWA, it’s used by around 30 researchers at KIST.