graphify v0.6.0 — SQL support: map your entire backend in one graph

[safishamsi]

graphify v0.6.0 is out

The big one for backend engineers.

SQL AST extraction

SQL files are now extracted deterministically via tree-sitter — no LLM, no tokens.

graphify maps:

• CREATE TABLE → table nodes
• CREATE VIEW → view nodes
• CREATE FUNCTION / CREATE PROCEDURE → function nodes
• REFERENCES constraints → foreign key edges
• FROM / JOIN clauses → reads_from edges

Your application code and your database schema are now in one graph. A FastAPI route that calls a PostgreSQL function that reads from a users table — that chain is visible. No more invisible backend logic.

Supports standard SQL, PostgreSQL, Snowflake, and any dialect tree-sitter-sql covers. Jinja-templated SQL (dbt, internal tooling) falls back gracefully.

pip install --upgrade graphifyy
pip install 'graphifyy[sql]'
/graphify .

YAML indexing

.yaml and .yml files are now indexed for semantic extraction automatically. Kubernetes manifests, Kustomize, Helm values, any config YAML — no setup needed.

What this unlocks

For the first time you can graph an entire backend stack in one run:

• Application code (Python, Go, TypeScript, ...)
• Database schema (.sql files)
• Infrastructure config (.yaml — k8s, docker-compose, Helm)
• Docs and architecture notes (.md, .pdf)

All in one graph. One query to understand the whole system.

Also in this release cycle

• Kimi K2.6 backend — pip install 'graphifyy[kimi]' + MOONSHOT_API_KEY for 3x cheaper semantic extraction
• Community labels now persist across re-clusters
• graphify update remembers your scan root
• Python 3.14+ supported

Full changelog: https://github.com/safishamsi/graphify/blob/v6/CHANGELOG.md
Release notes: https://github.com/safishamsi/graphify/releases/tag/v0.6.0

As always — if you hit issues, open one. If you build something cool with it, share it here.