graphtyy 0.1.0

AI coding assistant skill (Claude Code, CodeBuddy, Codex, OpenCode, Kilo Code, Cursor, Gemini CLI, Aider, OpenClaw, Factory Droid, Trae, Hermes, Kiro, Pi, Devin CLI, Google Antigravity) - turn any folder of code, docs, papers, images, or videos into a queryable knowledge graph

🇺🇸 English | 🇨🇳 简体中文 | 🇯🇵 日本語 | 🇰🇷 한국어 | 🇩🇪 Deutsch | 🇫🇷 Français | 🇪🇸 Español | 🇮🇳 हिन्दी | 🇧🇷 Português | 🇷🇺 Русский | 🇸🇦 العربية | 🇮🇹 Italiano | 🇵🇱 Polski | 🇳🇱 Nederlands | 🇹🇷 Türkçe | 🇺🇦 Українська | 🇻🇳 Tiếng Việt | 🇮🇩 Bahasa Indonesia | 🇸🇪 Svenska | 🇬🇷 Ελληνικά | 🇷🇴 Română | 🇨🇿 Čeština | 🇫🇮 Suomi | 🇩🇰 Dansk | 🇳🇴 Norsk | 🇭🇺 Magyar | 🇹🇭 ภาษาไทย | 🇺🇿 Oʻzbekcha | 🇹🇼 繁體中文 | 🇵🇭 Filipino

Graphty is an open-source fork of Graphify by Safi Shamsi, continued under the MIT license.

Stop grepping. Start asking. Type /graphty in your AI coding assistant and it maps your entire project — code, docs, PDFs, images, videos — into a knowledge graph your agent queries instead of reading files one at a time. Fewer tokens burned, fewer wrong turns, answers that span files.

Works in Claude Code, Codex, OpenCode, Kilo Code, Cursor, Gemini CLI, GitHub Copilot CLI, VS Code Copilot Chat, Aider, Amp, OpenClaw, Factory Droid, Trae, Hermes, Kimi Code, Kiro, Pi, Devin CLI, and Google Antigravity.

/graphty .

That's it. You get three files:

graphty-out/
├── graph.html       open in any browser — click nodes, filter, search
├── GRAPH_REPORT.md  the highlights: key concepts, surprising connections, suggested questions
└── graph.json       the full graph — query it anytime without re-reading your files

For a readable architecture page with Mermaid call-flow diagrams, run:

graphty export callflow-html

---

Why Graphty

• Works everywhere you code. 19+ AI coding tools supported — Claude Code, Codex, Cursor, Gemini, Copilot, and more — each a one-line graphty install.
• Query, don't grep. Your agent asks the graph a question and gets back a small, scoped subgraph instead of a wall of files. Run graphty benchmark to see the token savings on your own repo.
• Honest by design. Every edge is labeled EXTRACTED, INFERRED, or AMBIGUOUS, so you know what's certain and what to double-check — no hidden hallucinated links.
• Take the graph anywhere. Export to interactive HTML, JSON, Obsidian, Neo4j, GraphML, Graphviz DOT, and CSV, or get a one-glance overview with graphty stats (no LLM, instant).
• Persistent + incremental. The graph lives in graphty-out/ across sessions and updates on each commit via a git hook — graphty update . re-extracts only what changed (AST-only, no API cost).

Graphty is a community-maintained fork of Graphify by Safi Shamsi — the same extraction engine, kept actively maintained, with added exporters (DOT, CSV) and a fast stats command.

---

Prerequisites

Requirement	Minimum	Check	Install
Python	3.10+	python --version	python.org
uv (recommended)	any	uv --version	curl -LsSf https://astral.sh/uv/install.sh | sh
pipx (alternative)	any	pipx --version	pip install pipx

macOS quick install (Homebrew):

brew install python@3.12 uv

Windows quick install:

winget install astral-sh.uv

Ubuntu/Debian:

sudo apt install python3.12 python3-pip pipx
# or install uv:
curl -LsSf https://astral.sh/uv/install.sh | sh

---

Install

Package name: The PyPI package is graphtyy (double-y) — graphty was already taken, so this follows the same convention upstream used (graphifyy). The CLI command, import, and skill name are all just graphty. You can also install from source (Option A) if you prefer.

Step 1 — install the package:

# Option A — from source (works today):
git clone https://github.com/m3rcha/graphty.git
uv tool install ./graphty        # uv puts `graphty` on PATH automatically
# or: pipx install ./graphty

# Option B — from PyPI (once published):
uv tool install graphtyy
# Alternatives: pipx install graphtyy  /  pip install graphtyy (may need PATH setup)

Step 2 — register the skill with your AI assistant:

graphty install

That's it. Open your AI assistant and type /graphty .

To install the assistant skill into the current repository instead of your user profile, add --project:

graphty install --project
graphty install --project --platform codex

Project-scoped installs write under the current directory, for example .claude/skills/graphty/SKILL.md or .agents/skills/graphty/SKILL.md (plus a references/ sidecar the skill loads on demand), and print a git add hint for files that can be committed. Per-platform commands that support project-scoped installs accept the same flag, for example graphty claude install --project or graphty codex install --project.

PowerShell note: Use graphty . not /graphty . — the leading slash is a path separator in PowerShell.

graphty: command not found? Use uv tool install graphtyy or pipx install graphtyy — both put the CLI on PATH automatically. With plain pip, add ~/.local/bin (Linux) or ~/Library/Python/3.x/bin (Mac) to your PATH, or run python -m graphty.

Avoid pip install on Mac/Windows if possible. The skill resolves Python at runtime from graphty-out/.graphty_python; if that points to a different environment than where pip installed the package, you'll get ModuleNotFoundError: No module named 'graphty'. uv tool install and pipx install isolate the package in their own env and avoid this entirely.

Git hooks and uv tool / pipx: graphty hook install embeds the current interpreter path directly into the hook scripts at install time, so the post-commit hook fires correctly even in GUI git clients and CI runners where ~/.local/bin is not on PATH. If you reinstall or upgrade graphtyy, re-run graphty hook install to refresh the embedded path.

Pick your platform

Platform	Install command
Claude Code (Linux/Mac)	graphty install
Claude Code (Windows)	graphty install (auto-detected) or graphty install --platform windows
CodeBuddy	graphty install --platform codebuddy
Codex	graphty install --platform codex
OpenCode	graphty install --platform opencode
Kilo Code	graphty install --platform kilo
GitHub Copilot CLI	graphty install --platform copilot
VS Code Copilot Chat	graphty vscode install
Aider	graphty install --platform aider
OpenClaw	graphty install --platform claw
Factory Droid	graphty install --platform droid
Trae	graphty install --platform trae
Trae CN	graphty install --platform trae-cn
Gemini CLI	graphty install --platform gemini
Hermes	graphty install --platform hermes
Kimi Code	graphty install --platform kimi
Amp	graphty amp install
Kiro IDE/CLI	graphty kiro install
Pi coding agent	graphty install --platform pi
Cursor	graphty cursor install
Devin CLI	graphty devin install
Google Antigravity	graphty antigravity install

Codex users also need multi_agent = true under [features] in ~/.codex/config.toml for parallel extraction. CodeBuddy uses the same Agent tool and PreToolUse hook mechanism as Claude Code. Factory Droid uses the Task tool for parallel subagent dispatch. OpenClaw and Aider use sequential extraction (parallel agent support is still early on those platforms). Trae uses the Agent tool for parallel subagent dispatch and does not support PreToolUse hooks — AGENTS.md is the always-on mechanism.

Codex uses $graphty instead of /graphty.

Optional extras

Install only what you need:

Extra	What it adds	Install
pdf	PDF extraction	uv tool install "graphtyy[pdf]"
office	.docx and .xlsx support	uv tool install "graphtyy[office]"
google	Google Sheets rendering	uv tool install "graphtyy[google]"
video	Video/audio transcription (faster-whisper + yt-dlp)	uv tool install "graphtyy[video]"
mcp	MCP stdio server	uv tool install "graphtyy[mcp]"
neo4j	Neo4j push support	uv tool install "graphtyy[neo4j]"
svg	SVG graph export	uv tool install "graphtyy[svg]"
leiden	Leiden community detection (Python < 3.13 only)	uv tool install "graphtyy[leiden]"
ollama	Ollama local inference	uv tool install "graphtyy[ollama]"
openai	OpenAI / OpenAI-compatible APIs	uv tool install "graphtyy[openai]"
gemini	Google Gemini API	uv tool install "graphtyy[gemini]"
anthropic	Anthropic Claude API (--backend claude, uses ANTHROPIC_API_KEY)	uv tool install "graphtyy[anthropic]"
bedrock	AWS Bedrock (uses IAM, no API key)	uv tool install "graphtyy[bedrock]"
azure	Azure OpenAI Service (--backend azure, uses AZURE_OPENAI_API_KEY + AZURE_OPENAI_ENDPOINT)	uv tool install "graphtyy[openai]"
sql	SQL schema extraction	uv tool install "graphtyy[sql]"
postgres	Live PostgreSQL introspection (--postgres DSN)	uv tool install "graphtyy[postgres]"
dm	BYOND DreamMaker .dm/.dme AST extraction (may need a C compiler + python3-dev if no wheel matches your platform)	uv tool install "graphtyy[dm]"
terraform	Terraform / HCL .tf/.tfvars/.hcl AST extraction	uv tool install "graphtyy[terraform]"
chinese	Chinese query segmentation (jieba)	uv tool install "graphtyy[chinese]"
all	Everything above	uv tool install "graphtyy[all]"

---

Make your assistant always use the graph

Run this once in your project after building a graph:

Platform	Command
Claude Code	graphty claude install
CodeBuddy	graphty codebuddy install
Codex	graphty codex install
OpenCode	graphty opencode install
Kilo Code	graphty kilo install
GitHub Copilot CLI	graphty copilot install
VS Code Copilot Chat	graphty vscode install
Aider	graphty aider install
OpenClaw	graphty claw install
Factory Droid	graphty droid install
Trae	graphty trae install
Trae CN	graphty trae-cn install
Cursor	graphty cursor install
Gemini CLI	graphty gemini install
Hermes	graphty hermes install
Kimi Code	graphty install --platform kimi
Amp	graphty amp install
Kiro IDE/CLI	graphty kiro install
Pi coding agent	graphty pi install
Devin CLI	graphty devin install
Google Antigravity	graphty antigravity install

This writes a small config file that tells your assistant to consult the knowledge graph for codebase questions — preferring scoped queries like graphty query "<question>" over reading the full report or grepping raw files. On platforms that support payload-bearing hooks (Claude Code, Gemini CLI), a hook fires automatically before search-style tool calls (and, on Claude Code, before reading source files one by one via the Read/Glob tools) and nudges your assistant toward the graph path. On the others (Codex, OpenCode, Cursor, etc.), the persistent instruction files (AGENTS.md, .cursor/rules/, etc.) provide the same query-first guidance. GRAPH_REPORT.md is still available for broad architecture review.

CodeBuddy does the same two things as Claude Code: writes a CODEBUDDY.md section telling CodeBuddy to read graphty-out/GRAPH_REPORT.md before answering architecture questions, and installs PreToolUse hooks (.codebuddy/settings.json) that fire before Bash search commands and file reads, nudging toward graphty query instead.

Codex writes to AGENTS.md and also installs a PreToolUse hook in .codex/hooks.json that fires before every Bash tool call — same always-on mechanism as Claude Code.

To remove graphty from all platforms at once: graphty uninstall (add --purge to also delete graphty-out/). Or use the per-platform command (e.g. graphty claude uninstall).

---

Kilo Code installs the Graphty skill to ~/.config/kilo/skills/graphty/SKILL.md and a native /graphty command to ~/.config/kilo/command/graphty.md. graphty kilo install also writes AGENTS.md plus a native tool.execute.before plugin (.kilo/plugins/graphty.js + .kilo/kilo.json or .kilo/kilo.jsonc registration) so Kilo gets the same always-on graph reminder behavior through native .kilo config.

Cursor writes .cursor/rules/graphty.mdc with alwaysApply: true — Cursor includes it in every conversation automatically, no hook needed.

What's in the report

• God nodes — the most-connected concepts in your project. Everything flows through these.
• Surprising connections — links between things that live in different files or modules. Ranked by how unexpected they are.
• The "why" — inline comments (# NOTE:, # WHY:, # HACK:), docstrings, and design rationale from docs are extracted as separate nodes linked to the code they explain.
• Suggested questions — 4–5 questions the graph is uniquely positioned to answer.
• Confidence tags — every inferred relationship is marked EXTRACTED, INFERRED, or AMBIGUOUS. You always know what was found vs guessed.

---

What files it handles

Type	Extensions
Code (28 tree-sitter grammars)	.py .ts .js .jsx .tsx .mjs .go .rs .java .c .cpp .h .hpp .rb .cs .kt .scala .php .swift .lua .luau .zig .ps1 .ex .exs .m .mm .jl .vue .svelte .astro .groovy .gradle .dart .v .sv .svh .sql .f .f90 .f95 .f03 .f08 .pas .pp .dpr .dpk .lpr .inc .dfm .lfm .lpk .sh .bash .json .dm .dme .dmi .dmm .dmf .sln .slnx .csproj .fsproj .vbproj .razor .cshtml (.dm/.dme requires uv tool install graphtyy[dm])
Salesforce Apex	.cls .trigger (regex-based; classes, interfaces, enums, methods, triggers, SOQL/DML edges)
Terraform / HCL	.tf .tfvars .hcl (requires uv tool install graphtyy[terraform])
MCP configs	.mcp.json mcp.json mcp_servers.json claude_desktop_config.json — extracts server nodes, package refs, env var requirements
Docs	.md .mdx .qmd .html .txt .rst .yaml .yml
Office	.docx .xlsx (requires uv tool install graphtyy[office])
Google Workspace	.gdoc .gsheet .gslides (opt-in; requires gws auth and --google-workspace; Sheets need uv tool install graphtyy[google])
PDFs	.pdf
Images	.png .jpg .webp .gif
Video / Audio	.mp4 .mov .mp3 .wav and more (requires uv tool install graphtyy[video])
YouTube / URLs	any video URL (requires uv tool install graphtyy[video])

Code is extracted locally with no API calls (AST via tree-sitter). Everything else goes through your AI assistant's model API.

Google Drive for desktop .gdoc, .gsheet, and .gslides files are shortcut pointers, not document content. To include native Google Docs, Sheets, and Slides in a headless extraction, install and authenticate the gws CLI, then run:

uv tool install "graphtyy[google]"  # needed for Google Sheets table rendering
gws auth login -s drive
graphty extract ./docs --google-workspace

You can also set GRAPHTY_GOOGLE_WORKSPACE=1. Graphty exports shortcuts into graphty-out/converted/ as Markdown sidecars, then extracts those files.

---

Common commands

/graphty .                        # build graph for current folder
/graphty ./docs --update          # re-extract only changed files
/graphty . --cluster-only         # rerun clustering without re-extracting
/graphty . --cluster-only --resolution 1.5      # more granular communities
/graphty . --cluster-only --exclude-hubs 99     # suppress utility super-hubs from god-node rankings
/graphty . --no-viz               # skip the HTML, just the report + JSON
/graphty . --wiki                 # build a markdown wiki from the graph
graphty export callflow-html      # Mermaid architecture/call-flow HTML (auto-regenerates on every git commit if hook is installed)
graphty export dot                # Graphviz DOT (render with `dot -Tsvg graphty-out/graph.dot`)
graphty export csv                # nodes.csv + edges.csv for spreadsheets / pandas
graphty stats                     # quick terminal summary of the graph (no LLM)

/graphty query "what connects auth to the database?"
/graphty path "UserService" "DatabasePool"
/graphty explain "RateLimiter"

/graphty add https://arxiv.org/abs/1706.03762   # fetch a paper and add it
/graphty add <youtube-url>                       # transcribe and add a video

graphty hook install              # auto-rebuild on git commit
graphty merge-graphs a.json b.json              # combine two graphs

graphty prs                       # PR dashboard: CI state, review status, worktree mapping
graphty prs 42                    # deep dive on PR #42 with graph impact
graphty prs --triage              # AI ranks your review queue (uses whatever backend is configured)
graphty prs --conflicts           # PRs sharing graph communities — merge-order risk

See the full command reference below.

---

Ignoring files

Create a .graphtyignore in your project root — same syntax as .gitignore, including ! negation.

.gitignore is respected automatically. If no .graphtyignore is present in a directory, graphty falls back to the .gitignore in that directory. If both exist, .graphtyignore takes priority. Subdirectory scoping works the same way as git — an ignore file only affects its own subtree.

# .graphtyignore
node_modules/
dist/
*.generated.py

# only index src/, ignore everything else
*
!src/
!src/**

---

Team setup

graphty-out/ is meant to be committed to git so everyone on the team starts with a map.

Recommended .gitignore additions:

graphty-out/cost.json        # local only
# graphty-out/cache/         # optional: commit for speed, skip to keep repo small

manifest.json is now portable — keys are stored as relative paths and re-anchored on load, so committing it is safe and avoids a full rebuild on first checkout.

Workflow:

1. One person runs /graphty . and commits graphty-out/.
2. Everyone pulls — their assistant reads the graph immediately.
3. Run graphty hook install to auto-rebuild after each commit (AST only, no API cost). This also sets up a git merge driver so graph.json is never left with conflict markers — two devs committing in parallel get their graphs union-merged automatically.
4. When docs or papers change, run /graphty --update to refresh those nodes.

---

Using the graph directly

# query the graph from the terminal
graphty query "show the auth flow"
graphty query "what connects DigestAuth to Response?" --graph graphty-out/graph.json

# expose the graph as an MCP server (for repeated tool-call access)
python -m graphty.serve graphty-out/graph.json

# register with Kimi Code:
kimi mcp add --transport stdio graphty -- python -m graphty.serve graphty-out/graph.json

# or serve over HTTP so a whole team points at one URL (no local graphty needed):
python -m graphty.serve graphty-out/graph.json --transport http --port 8080
python -m graphty.serve graphty-out/graph.json --transport http --host 0.0.0.0 --api-key "$SECRET"

The MCP server gives your assistant structured access: query_graph, get_node, get_neighbors, shortest_path, list_prs, get_pr_impact, triage_prs.

Shared HTTP server

--transport stdio (the default) spawns one local server per developer. --transport http serves the same tools over the MCP Streamable HTTP transport, so a single shared process can serve the graph for the whole team — clients point their IDE MCP config at http://<host>:8080/mcp instead of running graphty locally.

Flag	Default	Purpose
--transport {stdio,http}	stdio	Transport to serve on
--host	127.0.0.1	HTTP bind host (use 0.0.0.0 to expose beyond localhost)
--port	8080	HTTP bind port
--api-key	env GRAPHTY_API_KEY	Require Authorization: Bearer <key> (or X-API-Key)
--path	/mcp	HTTP mount path
--json-response	off	Return plain JSON instead of SSE streams
--stateless	off	No per-session state (for load-balanced / CI deployments)
--session-timeout	3600	Reap idle stateful sessions after N seconds (0 disables)

The default 127.0.0.1 bind is loopback-only. Set --host 0.0.0.0 and --api-key together when exposing on a shared host. Run it in a container:

docker build -t graphty .
docker run -p 8080:8080 -v "$(pwd)/graphty-out:/data" graphty \
  /data/graph.json --transport http --host 0.0.0.0 --api-key "$SECRET"

WSL / Linux note: Ubuntu ships python3, not python. Use a venv to avoid conflicts:

python3 -m venv .venv && .venv/bin/pip install "graphtyy[mcp]"

---

Environment variables

These are only needed for headless / CI extraction (graphty extract). When running via the /graphty skill inside your IDE, the model API is provided by your IDE session — no extra keys needed.

Variable	Used for	When required
ANTHROPIC_API_KEY	Claude (Anthropic) backend	--backend claude
GEMINI_API_KEY or GOOGLE_API_KEY	Google Gemini backend	--backend gemini
OPENAI_API_KEY	OpenAI or OpenAI-compatible APIs	--backend openai
DEEPSEEK_API_KEY	DeepSeek backend	--backend deepseek
MOONSHOT_API_KEY	Kimi Code backend	--backend kimi
OLLAMA_BASE_URL	Ollama local inference URL	--backend ollama (default: http://localhost:11434)
OLLAMA_MODEL	Ollama model name	--backend ollama (default: auto-detect)
GRAPHTY_OLLAMA_NUM_CTX	Override Ollama KV-cache window size	optional — auto-sized by default
GRAPHTY_OLLAMA_KEEP_ALIVE	Minutes to keep Ollama model loaded	optional — set 0 to unload after each chunk
AZURE_OPENAI_API_KEY	Azure OpenAI Service backend	--backend azure
AZURE_OPENAI_ENDPOINT	Azure resource endpoint URL	--backend azure (required alongside API key)
AZURE_OPENAI_API_VERSION	Azure API version override	optional — default 2024-12-01-preview
AZURE_OPENAI_DEPLOYMENT or GRAPHTY_AZURE_MODEL	Azure deployment name	optional — default gpt-4o
AWS_* / ~/.aws/credentials	AWS Bedrock — standard credential chain	--backend bedrock (no API key, uses IAM)
GRAPHTY_MAX_WORKERS	AST parallelism thread count	optional — also --max-workers flag
GRAPHTY_MAX_OUTPUT_TOKENS	Raise output cap for dense corpora	optional — e.g. 32768 for large files
GRAPHTY_API_TIMEOUT	Per-call timeout in seconds for HTTP, claude-cli, and Anthropic SDK backends (default: 600)	optional — also --api-timeout flag
GRAPHTY_FORCE	Force graph rebuild even with fewer nodes	optional — also --force flag
GRAPHTY_GOOGLE_WORKSPACE	Auto-enable Google Workspace export	optional — set to 1
GRAPHTY_TRIAGE_BACKEND	Backend for graphty prs --triage	optional — auto-detected from available keys
GRAPHTY_TRIAGE_MODEL	Model override for triage	optional — e.g. claude-opus-4-7
GRAPHTY_QUERY_LOG	Override query log path (default: ~/.cache/graphty-queries.log)	optional — set to empty or /dev/null to silence
GRAPHTY_QUERY_LOG_DISABLE	Set to 1 to disable query logging entirely	optional
GRAPHTY_QUERY_LOG_RESPONSES	Set to 1 to also log full subgraph responses (off by default)	optional
GRAPHTY_MAX_GRAPH_BYTES	Override the 512 MiB graph.json size cap — e.g. 700MB, 2GB, or plain bytes	optional — useful for very large corpora
GRAPHTY_LLM_TEMPERATURE	Override LLM temperature for semantic extraction — e.g. 0.7, or none to omit	optional — auto-omitted for o1/o3/o4/gpt-5 reasoning models

---

Privacy

• Code files — processed locally via tree-sitter. Nothing leaves your machine. A code-only corpus requires no API key — graphty extract runs fully offline.
• Video / audio — transcribed locally with faster-whisper. Nothing leaves your machine.
• Docs, PDFs, images — sent to your AI assistant for semantic extraction (via the /graphty skill, using whatever model your IDE session runs). Headless graphty extract requires GEMINI_API_KEY / GOOGLE_API_KEY (Gemini), MOONSHOT_API_KEY (Kimi), ANTHROPIC_API_KEY (Claude), OPENAI_API_KEY (OpenAI), DEEPSEEK_API_KEY (DeepSeek), a running Ollama instance (OLLAMA_BASE_URL), AWS credentials via the standard provider chain (Bedrock - no API key needed, uses IAM), or the claude CLI binary (Claude Code - no API key needed, uses your Claude subscription). The --dedup-llm flag uses the same key.
• Data residency — graphty extract auto-detects which provider to use based on which API key is set (priority: Gemini → Kimi → Claude → OpenAI → DeepSeek → Azure → Bedrock → Ollama). For code with data-residency requirements, use --backend ollama (fully local) or pass an explicit --backend flag. Kimi (MOONSHOT_API_KEY) routes to Moonshot AI servers in China.
• No telemetry, no usage tracking, no analytics.
• Query logging — every graphty query, graphty path, graphty explain, and MCP query_graph call is logged to ~/.cache/graphty-queries.log in JSON Lines format (timestamp, question, corpus, nodes returned, duration). Full subgraph responses are not stored by default. Set GRAPHTY_QUERY_LOG_DISABLE=1 to opt out, or GRAPHTY_QUERY_LOG=/dev/null to silence without disabling the code path.

---

Troubleshooting

graphty: command not found after pip install graphtyy pip installs scripts to a user bin directory that may not be on your PATH. Fix:

• macOS: add ~/Library/Python/3.x/bin to your PATH in ~/.zshrc
• Linux: add ~/.local/bin to your PATH in ~/.bashrc
• Or use uv tool install graphtyy / pipx install graphtyy — both manage PATH automatically.

python -m graphty works but graphty command doesn't Your shell's PATH doesn't include the Python scripts directory. Use uv or pipx instead of plain pip.

/graphty . causes "path not recognized" in PowerShell PowerShell treats a leading / as a path separator. Use graphty . (no slash) on Windows.

Graph has fewer nodes after --update or rebuild If a refactor deleted files, the old nodes linger. Pass --force (or set GRAPHTY_FORCE=1) to overwrite even when the rebuild has fewer nodes.

Graph has duplicate nodes for the same entity (ghost duplicates) Ghost duplicates (same symbol appearing twice — once from AST extraction with a source location, once from semantic extraction without) are now automatically merged at build time. If you see this in a graph built before v0.8.33, run a full re-extract to clean up:

graphty extract . --force

Ollama runs out of VRAM / context window exceeded The KV-cache window is auto-sized but may be too large for your GPU. Reduce it:

GRAPHTY_OLLAMA_NUM_CTX=8192 graphty extract ./docs --backend ollama --token-budget 4000

Graph HTML is too large to open in a browser (>5000 nodes) Skip HTML generation and use the JSON directly:

graphty cluster-only ./my-project --no-viz
graphty query "..."

graph.json has conflict markers after two devs commit at once Run graphty hook install — it sets up a git merge driver that union-merges graph.json automatically so conflicts never happen.

Extraction returns empty nodes/edges for docs or PDFs Docs, PDFs, and images require an LLM call — code-only corpora need no key. Check that your API key is set and the backend is correct:

ANTHROPIC_API_KEY=sk-... graphty extract ./docs --backend claude

Skill version mismatch warning in your IDE Your installed graphty version is different from the skill file. Update:

uv tool upgrade graphtyy
graphty install  # overwrites the skill file

---

Full command reference

/graphty                          # run on current directory
/graphty ./raw                    # run on a specific folder
/graphty ./raw --mode deep        # more aggressive relationship extraction
/graphty ./raw --update           # re-extract only changed files
/graphty ./raw --directed         # preserve edge direction
/graphty ./raw --cluster-only     # rerun clustering on existing graph
/graphty ./raw --no-viz           # skip HTML visualization
/graphty ./raw --obsidian         # generate Obsidian vault
/graphty ./raw --wiki             # build agent-crawlable markdown wiki
/graphty ./raw --svg              # export graph.svg
/graphty ./raw --graphml          # export for Gephi / yEd
/graphty ./raw --neo4j            # generate cypher.txt for Neo4j
/graphty ./raw --neo4j-push bolt://localhost:7687
/graphty ./raw --watch            # auto-sync as files change
/graphty ./raw --mcp              # start MCP stdio server

/graphty add https://arxiv.org/abs/1706.03762
/graphty add <video-url>
/graphty add https://... --author "Name" --contributor "Name"

/graphty query "what connects attention to the optimizer?"
/graphty query "..." --dfs --budget 1500
/graphty path "DigestAuth" "Response"
/graphty explain "SwinTransformer"

graphty uninstall                 # remove from all platforms in one shot
graphty uninstall --purge         # also delete graphty-out/
graphty uninstall --project --platform codex  # remove project-scoped install files only

graphty hook install              # post-commit + post-checkout hooks
graphty hook uninstall
graphty hook status

# always-on assistant instructions - platform-specific
graphty claude install            # CLAUDE.md + PreToolUse hook (Claude Code)
graphty claude uninstall
graphty codebuddy install         # CODEBUDDY.md + PreToolUse hook (CodeBuddy)
graphty codebuddy uninstall
graphty codex install             # AGENTS.md + PreToolUse hook in .codex/hooks.json (Codex)
graphty opencode install          # AGENTS.md + tool.execute.before plugin (OpenCode)
graphty kilo install              # native Kilo skill + /graphty command + AGENTS.md + .kilo plugin
graphty kilo uninstall
graphty cursor install            # .cursor/rules/graphty.mdc (Cursor)
graphty cursor uninstall
graphty gemini install            # GEMINI.md + BeforeTool hook (Gemini CLI)
graphty gemini uninstall
graphty copilot install           # skill file (GitHub Copilot CLI)
graphty copilot uninstall
graphty aider install             # AGENTS.md (Aider)
graphty aider uninstall
graphty claw install              # AGENTS.md (OpenClaw)
graphty claw uninstall
graphty droid install             # AGENTS.md (Factory Droid)
graphty droid uninstall
graphty trae install              # AGENTS.md (Trae)
graphty trae uninstall
graphty trae-cn install           # AGENTS.md (Trae CN)
graphty trae-cn uninstall
graphty hermes install             # AGENTS.md + ~/.hermes/skills/ (Hermes)
graphty hermes uninstall
graphty amp install               # skill file (Amp)
graphty amp uninstall
graphty kiro install               # .kiro/skills/ + .kiro/steering/graphty.md (Kiro IDE/CLI)
graphty kiro uninstall
graphty pi install                # skill file (Pi coding agent)
graphty pi uninstall
graphty devin install             # skill file + .windsurf/rules/graphty.md (Devin CLI)
graphty devin uninstall
graphty antigravity install       # .agents/rules + .agents/workflows (Google Antigravity)
graphty antigravity uninstall

graphty extract ./docs                        # headless LLM extraction for CI (no IDE needed)
graphty extract ./docs --backend gemini       # explicit backend: gemini, kimi, claude, openai, deepseek, ollama, bedrock, or claude-cli
graphty extract ./docs --backend gemini --model gemini-3.1-pro-preview
graphty extract ./docs --backend ollama       # local Ollama (set OLLAMA_BASE_URL / OLLAMA_MODEL) - no API key needed for loopback
GRAPHTY_OLLAMA_NUM_CTX=32768 graphty extract ./docs --backend ollama   # override KV-cache window (auto-sized by default)
GRAPHTY_OLLAMA_KEEP_ALIVE=0 graphty extract ./docs --backend ollama    # unload model after each chunk (saves VRAM on small GPUs)
graphty extract ./docs --backend bedrock      # AWS Bedrock via IAM - no API key, uses AWS credential chain
graphty extract ./docs --backend claude-cli   # route through Claude Code CLI - no API key, uses your Claude subscription
graphty extract ./docs --backend azure        # Azure OpenAI (set AZURE_OPENAI_API_KEY + AZURE_OPENAI_ENDPOINT)
graphty extract ./docs --max-workers 16       # AST parallelism (also GRAPHTY_MAX_WORKERS)
graphty extract --postgres "postgresql://user:pass@host/db"   # introspect live PostgreSQL schema directly
graphty extract ./docs --token-budget 30000   # smaller semantic chunks for local/small models
graphty extract ./docs --max-concurrency 2    # fewer parallel LLM calls (useful for local inference)
graphty extract ./docs --api-timeout 900      # longer HTTP timeout for slow local models (default 600s)
graphty extract ./docs --google-workspace     # export .gdoc/.gsheet/.gslides via gws before extraction
graphty extract ./docs --mode deep            # richer semantic extraction via extended system prompt
graphty extract ./docs --no-cluster           # raw extraction only, skip clustering
graphty extract ./docs --force                # overwrite graph.json even if new graph has fewer nodes (use after refactors or to clear ghost duplicates)
graphty extract ./docs --dedup-llm            # LLM tiebreaker for ambiguous entity pairs (uses same API key)
graphty extract ./docs --global --as myrepo   # extract and register into the cross-project global graph
GRAPHTY_MAX_OUTPUT_TOKENS=32768 graphty extract ./docs --backend claude  # raise output cap for dense corpora

graphty export callflow-html                       # graphty-out/<project>-callflow.html
graphty export callflow-html --max-sections 8      # cap generated architecture sections
graphty export callflow-html --output docs/arch.html
graphty export callflow-html ./some-repo/graphty-out

graphty global add graphty-out/graph.json myrepo   # register a project graph into ~/.graphty/global.json
graphty global remove myrepo                         # remove a project from the global graph
graphty global list                                  # show all registered repos + node/edge counts
graphty global path                                  # print path to the global graph file

graphty prs                              # PR dashboard: CI, review, worktree, graph impact
graphty prs 42                           # deep dive on PR #42
graphty prs --triage                     # AI triage ranking (auto-detects backend from env)
graphty prs --worktrees                  # worktree → branch → PR mapping
graphty prs --conflicts                  # PRs sharing graph communities (merge-order risk)
graphty prs --base main                  # filter to PRs targeting a specific base branch
graphty prs --repo owner/repo            # run against a different GitHub repo
GRAPHTY_TRIAGE_BACKEND=kimi graphty prs --triage   # use a specific backend for triage

graphty clone https://github.com/karpathy/nanoGPT
graphty merge-graphs a.json b.json --out merged.json
graphty --version                                    # print installed version
graphty watch ./src
graphty check-update ./src
graphty update ./src
graphty update ./src --no-cluster  # skip reclustering, write raw AST graph only
graphty update ./src --force       # overwrite even if new graph has fewer nodes
graphty cluster-only ./my-project
graphty cluster-only ./my-project --graph path/to/graph.json  # custom graph location
graphty cluster-only ./my-project --resolution 1.5            # more, smaller communities
graphty cluster-only ./my-project --exclude-hubs 99           # exclude p99 degree nodes from partitioning
graphty cluster-only ./my-project --no-label                  # keep "Community N" placeholders
graphty cluster-only ./my-project --backend=gemini            # backend for community naming
graphty label ./my-project                                    # (re)name communities with the configured backend
graphty label ./my-project --backend=openai                   # force a specific backend

Community names: inside an agent (Claude Code, Gemini CLI) the agent names communities itself. When you run the bare CLI, cluster-only auto-names them with the configured backend (built-in or custom OpenAI-compatible provider) — pass --no-label to keep Community N, or run graphty label to (re)generate names on demand.

---

Learn more

• How it works — the extraction pipeline, community detection, confidence scoring, benchmarks
• ARCHITECTURE.md — module breakdown, how to add a language
• Optional integrations — Docker MCP Toolkit + SQLite

---

Contributing

Development setup

The project uses uv for dev workflow. Install it once, then:

git clone https://github.com/m3rcha/graphty.git
cd graphty
git checkout v8                        # active development branch

# Create the project venv and install graphty + all extras + the dev group
# (pytest). uv installs the dev dependency group by default; pass --no-dev to
# skip it.
uv sync --all-extras

Verify the editable install:

uv run graphty --version
uv run python -c "import graphty; print(graphty.__file__)"

Running tests

uv run pytest tests/ -q                # run the full suite
uv run pytest tests/test_extract.py -q # one module
uv run pytest tests/ -q -k "python"    # filter by name

macOS note: the test suite includes both sample.f90 and sample.F90 fixtures. These collide on case-insensitive HFS+ / APFS file systems. Run on Linux or in a Docker container if you need to test both Fortran variants simultaneously.

Git workflow

• Active development happens on the v8 branch.
• Commit style: fix: <description> / feat: <description> / docs: <description>
• Before opening a PR, run uv run pytest tests/ -q and confirm it passes.
• Add a fixture file to tests/fixtures/ and tests to tests/test_languages.py for any new language extractor.

What to contribute

Worked examples are the most useful contribution. Run /graphty on a real corpus, save the output to worked/{slug}/, write an honest review.md covering what the graph got right and wrong, and open a PR.

Extraction bugs — open an issue with the input file, the cache entry (graphty-out/cache/), and what was missed or wrong.

See ARCHITECTURE.md for module responsibilities and how to add a language.