Code Cartography

Code Cartography means Altretta ingests a codebase into the same signed graph as your notes and understands it structurally — every symbol, every cross-file reference, and everything a change would touch.

Your second brain already holds your notes, decisions, and docs. Code Cartography lets it hold your code too — not as flat text, but as a real graph of definitions and relationships that Altretta builds from the source itself.

This is the differentiator. Obsidian and Logseq treat a code file as a wall of text to search. Altretta parses it — extracting an AST symbol graph across ~20 languages — so it knows that this function calls that one, that this type is used across five files, and what would break if you changed it. No other PKM tool understands source code as a graph.

Why this matters

A codebase is a graph whether or not your tools admit it: functions call functions, types flow across files, decisions constrain modules. When that graph is invisible, you keep it in your head and it rots. Code Cartography makes it explicit and durable:

Structure, not string-matching. Altretta resolves references across files, so "who uses this?" is an answer, not a grep.
One graph for code and context. Symbols live beside the notes and decisions that explain them, so an architectural decision and the code it governs are connected — and can be checked against each other.
Signed and provenanced. Because code enters the same signed graph as your notes, the git provenance of every ingest is recorded in the signed DAG. You can prove which commit produced which structure.

What you can do with it

Ingest a codebase

Point Altretta at a repository and build an AST symbol graph across ~20 languages, recorded with its git provenance.

The symbol graph

Browse the top symbols and navigate the codebase structurally by its real cross-file relationships.

Impact Radius

For any symbol or change, see the blast radius across the graph — what breaks if you change this.

Doc & decision drift

Get warned when your docs or recorded decisions have drifted away from the current code.

Code Q&A

Ask your own AI about the codebase over MCP and get answers grounded in — and cited to — real source.

It inherits everything else

Because code lives in the same signed graph, the rest of Altretta applies to it unchanged:

Grounded, cited answers work over code too — answers quote real source lines instead of guessing.
Signed history covers every ingest, so the evolution of your codebase's structure is a verifiable record, not a mutable log.

New here? Start with Ingest a codebase, then explore The symbol graph. Impact Radius is where it earns its keep.

Grounded, cited answers

The trust model your AI uses to answer from real source — code or notes.

The semantic graph

Code symbols live in the same semantic graph that powers backlinks and structure over your notes.

Signed history

Every ingest is recorded with its git provenance in the signed DAG.

Glossary

Symbol graph, DAG, provenance, and the other terms used here, defined.