graphify v0.5.0 — clone any GitHub repo, cross-repo graphs, and data-loss protection

[safishamsi]

Hey everyone 👋

v0.5.0 is out on PyPI.

pip install --upgrade graphifyy

---

What's new

Clone any GitHub repo directly

graphify clone https://github.com/karpathy/nanoGPT

Clones to ~/.graphify/repos/karpathy/nanoGPT, runs the full pipeline on it. Repeat calls just git pull. Supports --branch and --out.

Cross-repo knowledge graphs
Run graphify on two repos, then merge their graphs into one:

graphify merge-graphs repo1/graphify-out/graph.json repo2/graphify-out/graph.json

Every node carries a repo tag so you can filter by origin. Useful for mapping dependencies across microservices or understanding how two projects relate.

Data-loss protection on --update
graphify now refuses to overwrite graph.json with a smaller graph — you get a clear warning instead of silent data loss. New build_merge() library function for safe incremental updates that only ever grows the graph.

Duplicate node deduplication
Parallel semantic subagents sometimes produced duplicate nodes for the same entity (achille_varzi and achille_varzi_c4). The subagent prompt now blocks chunk-suffix contamination at the source, and a post-merge deduplication pass catches any stragglers.

Bug fixes

• CLAUDE_CONFIG_DIR env var now respected on install (Issue: do not respect CLAUDE_CONFIG_DIR #527)
• graphify-out/ excluded from source scanning — own output no longer triggers false incremental refresh (detect_incremental() treats generated graphify-out docs as source input and creates false semantic-refresh pressure #524)

---

As always, open issues or drop questions in Q&A. Appreciate everyone who filed detailed bug reports — #479 and #451 especially.

[mdizo]

Does graphify clone add to my project's knowledge graph?