AnaxiGraph

Repository composition

Languages

Areas and their subsystems

Architecture map

Loading the best available architecture map.

Things worth checking

Current code findings

Files saved in AnaxiIndex

File inventory

One row per file in the selected scan. Filter and sort facts read from code, repository area, purpose, Git changes, test coverage, and what may deserve attention. Expand a row to see where each explanation came from and why a pattern may fit.

Fact Interpretation Review signal

Find file or purpose Area Subsystem Language Include docs, config & assets Rows

Attention only decides which files appear first. It uses code lines, total branches, direct code links, Git changes, and active findings. It is not a code-quality or pattern-fit grade.

			Pattern idea

Overlay Circle size Find file External libraries and packages Repository area boxes

Areas · all

Graph height

Wheel to zoom · drag to pan · select a node to inspect

From finding to checked result

How the review workflow works

A finding says what AnaxiGraph observed and why it may matter; it does not order you to change code. Read it, select it when you want an agent to act, then use a later scan to check whether the observed condition disappeared.

Short attention list · complete record

Architecture findings

The attention list keeps the most useful findings short. The complete record keeps every observation for inspection and filtering.

View

Saved Git history

How the code map changed over time

One architecture adviser for people and coding agents

Ask where to build or how to improve the design

AnaxiGraph does not edit code. It combines the current Charter, module meanings, code links, history, patterns, findings, and tests into the same recommendation here and through AnaxiMCP.

ANAXIGRAPH_GUIDE → ANAXIGRAPH_FILE / ANAXIGRAPH_IMPACT → refresh and ask again

Evidence-backed architecture guidance

What do you want to build or improve?

Describe the outcome. The answer recommends whether to reuse, extend, split, consolidate, move, delete, create, or retain code—and explains why.

Check before editing shared code

What could this change affect?

Enter a file or unique code name to see code that uses it directly or indirectly, relevant tests, possible database changes, files marked for extra care, and changes on other branches.

Likely task files, code that may be affected, or a planned finding handoff will appear here.

Repository connections

One shared model for people and coding agents

Repositories stay read-only. A scan turns their files, Git history, direct code links, findings, and optional test coverage into AnaxiIndex records. The dashboard, REST API, and AnaxiMCP all use that same living architecture model.

Read-only repository→Scanner→ AnaxiIndex→Dashboard · REST · AnaxiMCP

Live service configuration

Connected repositories

Add another repository

Initialize it beside the code

Run one non-destructive command from the repository you want to understand. It detects obvious areas and writes an editable policy plus a dedicated, separate read-only Docker service. Existing files are kept unless you explicitly use --force.

Create and start from the target repository
```
uvx anaxigraph init . --start
```
Open its dashboard
```
http://127.0.0.1:8765
```

Docker guide →

The Docker guide also covers an experimental multi-repository registry for one trusted operator. Keep it on loopback or behind SSH, never on an untrusted or shared network.

Coding-agent access

Use the same intelligence while coding

Run the command below in a normal terminal on the machine where Codex runs—not inside a Codex chat. Connect Codex or another MCP client to . Use ANAXIGRAPH_REPOSITORIES to choose a repository, then query findings, file details, change impact, or a goal-specific list of files and facts. AnaxiMCP reads AnaxiIndex; it does not grant the agent write access to the mounted repository.

AI-created code map

Build and refresh plain-language code descriptions

With semantic.provider: agent, the connected coding agent reads one saved task at a time and writes a checked description to AnaxiIndex using its own model and tokens. Other AI providers use a separate worker. Every path is opt-in, and the mounted repository remains read-only.

AI mapping setup →