pinned to function context
NotesResearch, captures, clips ExamplesSamples in your voice and prior outputs RulesConstraints and decision criteria Standing contextPortfolios, clients, trips, goals
how pachinko works
Build a Function. Repeatable calculations that think like you.
A Function is a set of saved instructions plus pinned context. Feed it notes, links, PDFs, bookmarks, or scheduled sources. Pachinko writes the result back into your work to move it forward.
Pin your context.
Attach notes, sections, examples, rules, preferences, and projects. These are the facts and constraints each run starts from.
Define the process.
Tell the Function how to process new information, leverage your pinned context, and what kind of output to produce.
Analyze how the provided news affects my stock portfolio. For each holding, explain the likely impact, reasoning, and any action I should consider.
Run new content.
Run selected notes by hand, or let Feeds pass in articles, PDFs, emails, bookmarks, transcripts, and other scheduled sources.
content in
fx Calculate portfolio impact
work out
Hand off the results as notes and tasks.
The result becomes editable work: a note, task, draft, warning, revision, or next step you can use immediately.
Call Functions from outside Pachinko.
MCP lets tools like Claude Code, Codex, and Cursor call your saved Functions, read visible project material, create sections and notes, and write output back into Pachinko.
MCP / external control
Claude Code, Codex, Cursor, and any MCP client can run saved Functions, read visible project material, create sections and notes, and write output back.
$ claude "run my health tracking function on blood-report.pdf" -> reading pachinko functions... -> converting PDF to markdown... -> initiating health tracking function