AI tool

Graphify

Visit website

Build a queryable code knowledge graph with explicit versus inferred relationships for coding assistants.

Starting price
Local code mapping: no API fee
Free access
Open-source CLI
Billing unit
Semantic models / Cloud separate
Platform
Python CLI and assistant integrations
Verified
26 Oct 2026

Overview

What you provide

  • A local repository, optionally with documentation and supported media.

What you get

  • An interactive graph, a Markdown report and a JSON graph for scoped queries.

Choosing this tool

Graphify is useful when a change spans several modules and repeated file searches obscure the dependency path. Use its local code-only mode for a bounded first evaluation. Aider has a repository map inside its coding workflow; Repomix packages source for model context without providing the same traversable graph.

A decision worked through

Before changing authentication, trace the path from a handler to its data access node and read the linked source. An INFERRED edge is a lead to verify, not proof of a runtime call. Refresh the graph after source changes so an old map does not become an authority over current code.

Graphify: alternatives and trade-offs

Compare Graphify on the documented workflow and cost boundary below; this is an official-source comparison, not a hands-on benchmark.

ProductPrice / unitFree allowanceInput → outputKey limit
GraphifyRead profileOfficial source ↗Local CLI free; model costs extraCode-only extractionRepository → queryable graphInferred edges need source verification
AiderRead profileOfficial source ↗Apache-2.0; model usage separateOpen coding clientRepository map → coding contextMap is selected to a token budget
RepomixOfficial source ↗MIT source-packing toolRepository packagingSource files → model-ready bundlePacked context is not a graph

Setup & workflow

  1. Install graphifyy from the official package reference.
  2. Register the assistant integration and build a code-only graph first.
  3. Use query, path or explain and inspect the cited source locations.
Requirements & installation
  • Python 3.10 or newer and uv or pipx; select an explicit model backend only if semantic extraction is needed.
uv tool install graphifyy

Limits & review

  • The official Python package is graphifyy, with two y characters. Local code extraction and cloud or semantic processing have different privacy and cost boundaries; CLI installation does not establish an MCP connection.

Your part

  • Check important graph edges against source code before approving a change.
Sources & verification3

This profile is based on official sources, not a hands-on product test.

Vendor descriptions and demos document advertised features. Editorial guidance is based on these sources.