codegraphcontext 0.5.1__tar.gz → 0.5.3__tar.gz

{codegraphcontext-0.5.1/src/codegraphcontext.egg-info → codegraphcontext-0.5.3}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: codegraphcontext
-Version: 0.5.1
+Version: 0.5.3
 Summary: An MCP server that indexes local code into a graph database to provide context to AI assistants.
 Author-email: Shashank Shekhar Singh <shashankshekharsingh1205@gmail.com>
 License: MIT License
@@ -25,9 +25,14 @@ License: MIT License
         OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
         SOFTWARE.
         
-Project-URL: Homepage, https://github.com/Shashankss1205/CodeGraphContext
-Project-URL: Bug Tracker, https://github.com/Shashankss1205/CodeGraphContext/issues
+Project-URL: Homepage, https://github.com/CodeGraphContext/CodeGraphContext
+Project-URL: Bug Tracker, https://github.com/CodeGraphContext/CodeGraphContext/issues
 Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3.10
+Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3.12
+Classifier: Programming Language :: Python :: 3.13
+Classifier: Programming Language :: Python :: 3.14
 Classifier: License :: OSI Approved :: MIT License
 Classifier: Operating System :: OS Independent
 Classifier: Development Status :: 3 - Alpha
@@ -43,9 +48,9 @@ Requires-Dist: typer>=0.9.0
 Requires-Dist: rich>=13.7.0
 Requires-Dist: inquirerpy>=0.3.4
 Requires-Dist: python-dotenv>=1.0.0
-Requires-Dist: tree-sitter<0.26.0,>=0.21.0; python_version != "3.13"
-Requires-Dist: tree-sitter-language-pack<1.0.0,>=0.6.0; python_version != "3.13"
-Requires-Dist: tree-sitter-c-sharp>=0.21.0; python_version != "3.13"
+Requires-Dist: tree-sitter<0.26.0,>=0.24.0
+Requires-Dist: tree-sitter-language-pack<2.0,>=1.6
+Requires-Dist: tree-sitter-c-sharp>=0.21.0
 Requires-Dist: pyyaml
 Requires-Dist: nbformat
 Requires-Dist: nbconvert>=7.16.6
@@ -55,15 +60,15 @@ Requires-Dist: falkordb<1.6,>=1.0
 Requires-Dist: falkordblite<0.10,>=0.7; sys_platform != "win32" and python_version >= "3.12"
 Requires-Dist: requests>=2.28.0
 Requires-Dist: protobuf<3.21,>=3.20
-Requires-Dist: kuzu; sys_platform == "win32" or (sys_platform != "win32" and python_version >= "3.10")
+Requires-Dist: kuzu; python_version < "3.14" or sys_platform == "linux"
 Requires-Dist: ladybug; sys_platform == "win32" or (sys_platform != "win32" and python_version >= "3.10")
 Requires-Dist: mcp>=1.0.0
 Requires-Dist: fastapi>=0.100.0
 Requires-Dist: uvicorn>=0.22.0
 Provides-Extra: parsing
-Requires-Dist: tree-sitter<0.26.0,>=0.21.0; python_version != "3.13" and extra == "parsing"
-Requires-Dist: tree-sitter-language-pack<1.0.0,>=0.6.0; python_version != "3.13" and extra == "parsing"
-Requires-Dist: tree-sitter-c-sharp>=0.21.0; python_version != "3.13" and extra == "parsing"
+Requires-Dist: tree-sitter<0.26.0,>=0.24.0; extra == "parsing"
+Requires-Dist: tree-sitter-language-pack<2.0,>=1.6; extra == "parsing"
+Requires-Dist: tree-sitter-c-sharp>=0.21.0; extra == "parsing"
 Provides-Extra: falkordb-embedded
 Requires-Dist: falkordblite<0.10,>=0.7; (sys_platform != "win32" and python_version >= "3.12") and extra == "falkordb-embedded"
 Requires-Dist: falkordb<1.6,>=1.0; extra == "falkordb-embedded"
@@ -72,11 +77,13 @@ Provides-Extra: falkordb-remote
 Requires-Dist: falkordb<1.6,>=1.0; extra == "falkordb-remote"
 Requires-Dist: redis<6,>=5; extra == "falkordb-remote"
 Provides-Extra: kuzu
-Requires-Dist: kuzu; (sys_platform == "win32" or (sys_platform != "win32" and python_version >= "3.10")) and extra == "kuzu"
+Requires-Dist: kuzu; (python_version < "3.14" or sys_platform == "linux") and extra == "kuzu"
 Provides-Extra: ladybug
 Requires-Dist: ladybug; (sys_platform == "win32" or (sys_platform != "win32" and python_version >= "3.10")) and extra == "ladybug"
 Provides-Extra: neo4j
 Requires-Dist: neo4j>=5.15.0; extra == "neo4j"
+Provides-Extra: gcf
+Requires-Dist: gcf-python>=1.0.0; extra == "gcf"
 Provides-Extra: dev
 Requires-Dist: pytest>=7.4.0; extra == "dev"
 Requires-Dist: black>=23.11.0; extra == "dev"
@@ -102,7 +109,7 @@ Dynamic: license-file
 - 🇯🇵 [日本語](docs/translations/README.ja.md)
 - 🇪🇸 Español (Soon)
 
-🌍 **Help translate CodeGraphContext to your language by raising an issue & PR on [GitHub Issues](https://github.com/Shashankss1205/CodeGraphContext/issues)!**
+🌍 **Help translate CodeGraphContext to your language by raising an issue & PR on [GitHub Issues](https://github.com/CodeGraphContext/CodeGraphContext/issues)!**
 
 <p align="center">
   <br>
@@ -190,6 +197,43 @@ A powerful **MCP server** and **CLI toolkit** that indexes local code into a gra
 
 ---
 
+## Why CodeGraphContext?
+
+CodeGraphContext is built for the moment when plain text search stops being enough. It turns a repository into a graph of files, symbols, calls, inheritance, imports, and relationships so you can move from "where is this defined?" to "how does this actually connect?" without jumping between tools.
+
+```mermaid
+flowchart LR
+  A[Source code] --> B[Tree-sitter and SCIP indexers]
+  B --> C[Graph database]
+  C --> D[CLI queries]
+  C --> E[MCP server]
+  D --> F[Direct analysis]
+  E --> G[AI assistants with repo context]
+```
+
+### Comparison at a glance
+
+| Approach | Best for | Tradeoff |
+| --- | --- | --- |
+| `grep` / file search | Exact string lookup | Misses relationships and code structure |
+| RAG over code chunks | Natural-language retrieval | Can lose symbol-level precision |
+| CodeGraphContext | Repository-wide reasoning | Requires an indexing step |
+
+### Common use cases
+
+- Trace call chains across files and packages.
+- Understand class hierarchies, imports, and module boundaries.
+- Find dead code, hotspots, and unexpected dependencies.
+- Give an AI assistant the same structural context a maintainer would use.
+
+### FAQ
+
+- **Do I need a remote service?** No. CGC works locally with embedded backends or with external graph databases when you want them.
+- **Is it only for one language?** No. It supports a broad mix of mainstream languages and keeps expanding.
+- **Can I use it with an AI assistant?** Yes. That is one of the main reasons the MCP server exists.
+
+---
+
 ## Project Details
 - **Version:** 0.5.1
 - **Authors:** Shashank Shekhar Singh <shashankshekharsingh1205@gmail.com>
@@ -227,6 +271,49 @@ A powerful **MCP server** and **CLI toolkit** that indexes local code into a gra
 -   **Flexible Database Backend:** FalkorDB Lite (Default), KuzuDB, LadybugDB, FalkorDB Remote, Nornic DB, or Neo4j (all platforms via Docker/native).
 
 
+---
+## Architecture Overview
+
+CodeGraphContext transforms source code into a queryable knowledge graph that can be explored through the CLI or AI assistants via MCP.
+
+```mermaid
+flowchart TD
+    A[Code Repository] --> B[Tree-sitter / SCIP Indexing]
+    B --> C[Knowledge Graph]
+    C --> D[Graph Database]
+    D --> E[CLI Toolkit]
+    D --> F[MCP Server]
+    F --> G[AI Assistant]
+```
+
+### Workflow
+
+1. Source code is parsed using Tree-sitter or SCIP indexers.
+2. Relationships between functions, classes, imports, and calls are extracted.
+3. A knowledge graph is generated and stored in a graph database.
+4. Users can query the graph through the CLI or AI assistants using MCP.
+
+## 🏗️ Architecture & Workflow
+
+CodeGraphContext parses your source code and builds a comprehensive knowledge graph. This graph can be queried directly via the CLI toolkit or exposed to AI assistants through the MCP server.
+
+```mermaid
+flowchart TD
+    A[Code Repository] -->|Parsed by| B[Tree-sitter / SCIP Indexing]
+    B -->|Generates| C[Knowledge Graph]
+    C -->|Stored in| D[(Graph Database)]
+    D -->|Queried via| E[CLI Toolkit]
+    D -->|Served by| F[MCP Server]
+    F -->|Provides context to| G[🤖 AI Assistant]
+
+    classDef default fill:#1f2937,stroke:#8b5cf6,stroke-width:2px,color:#fff;
+    classDef db fill:#0f172a,stroke:#3b82f6,stroke-width:2px,color:#fff;
+    classDef ai fill:#312e81,stroke:#a855f7,stroke-width:2px,color:#fff;
+    
+    D:::db
+    G:::ai
+```
+
 ---
 
 ## Supported Programming Languages
@@ -297,8 +384,8 @@ _If you’re using CodeGraphContext in your project, feel free to open a PR and
 - `rich>=13.7.0`
 - `inquirerpy>=0.3.4`
 - `python-dotenv>=1.0.0`
-- `tree-sitter>=0.21.0` (not installed on Python 3.13)
-- `tree-sitter-language-pack>=0.6.0` (not installed on Python 3.13)
+- `tree-sitter>=0.24.0,<0.26.0`
+- `tree-sitter-language-pack>=1.6,<2.0`
 - `pyyaml`
 - `pathspec>=0.12.1`
 - `falkordb>=1.0,<1.6`
@@ -332,6 +419,87 @@ _If you’re using CodeGraphContext in your project, feel free to open a PR and
     - **KuzuDB:** Cross-platform embedded backend.
     - **Neo4j:** Run `codegraphcontext neo4j setup` to use an external server.
 
+### 🧪 Local Development Setup
+
+If you want to run CodeGraphContext from a local clone and contribute changes, use an editable install:
+
+```bash
+git clone https://github.com/CodeGraphContext/CodeGraphContext.git
+cd CodeGraphContext
+python -m venv .venv
+source .venv/bin/activate
+pip install -e ".[dev]"
+```
+
+Then verify the install and try the local commands:
+
+```bash
+cgc --help
+cgc doctor
+cgc index .
+cgc analyze callers main
+```
+
+For MCP server work, finish the setup with:
+
+```bash
+cgc mcp setup
+cgc mcp start
+```
+
+If you are only installing the published package, `pip install codegraphcontext` is enough. If you are developing parser-heavy features and want the optional parsing extras too, install with `pip install -e ".[dev,parsing]"`.
+
+---
+
+## 🐳 Docker (No Python Required)
+
+Run CodeGraphContext without installing Python — pull the image and start indexing:
+
+### Quick Start
+```bash
+# Pull the latest image
+docker pull codegraphcontext/codegraphcontext:latest
+
+# Index your current directory
+docker run --rm -v "$(pwd):/workspace" codegraphcontext/codegraphcontext cgc index .
+
+# List indexed repos
+docker run --rm -v cgc-data:/home/cgc/.codegraphcontext \
+    codegraphcontext/codegraphcontext cgc list
+
+# Interactive shell
+docker run -it --rm -v "$(pwd):/workspace" \
+    -v cgc-data:/home/cgc/.codegraphcontext \
+    codegraphcontext/codegraphcontext bash
+```
+
+### Docker Compose
+```bash
+# Clone the repo and start with Docker Compose
+git clone https://github.com/CodeGraphContext/CodeGraphContext.git
+cd CodeGraphContext
+
+# Start with embedded DB
+docker compose run --rm cgc index .
+
+# Start with external FalkorDB
+docker compose --profile falkordb up -d
+
+# Start with visualization server
+docker compose --profile viz up -d
+# Open http://localhost:8080
+```
+
+### Available Docker Tags
+| Tag | Description |
+|---|---|
+| `latest` | Latest stable release |
+| `edge` | Latest from `main` branch (may be unstable) |
+| `0.4.19` | Specific version |
+| `0.4` | Latest patch in 0.4.x |
+
+For more advanced Docker usage, including running the MCP Server or connecting external databases, see our [Comprehensive Docker Guide](docs/DOCKER.md).
+
 ---
 
 ## 📋 Prerequisites
@@ -355,7 +523,7 @@ python --version
 ### Step 1: Install CodeGraphContext
 
 ```bash
-pip install codegraphcontext
+pip install codegraphcontext  # Installs CodeGraphContext and required dependencies
 ```
 
 This command installs CodeGraphContext and all required dependencies.
@@ -379,7 +547,7 @@ CodeGraphContext automatically uses an embedded database by default, so no addit
 ### Index a Repository
 
 ```bash
-codegraphcontext index .
+codegraphcontext index .  # Scans current directory and builds code graph
 ```
 
 This scans the current project and creates a searchable code graph.
@@ -407,7 +575,7 @@ Finds potentially unused code in the indexed repository.
 After indexing a repository, run:
 
 ```bash
-codegraphcontext list
+codegraphcontext list  # Shows all indexed repositories in the system
 ```
 
 If the command executes successfully and displays indexed repositories, your setup is complete and CodeGraphContext is ready to use.
@@ -536,11 +704,6 @@ Add the following server configuration to your client's settings file (e.g., VS
         "mcp",
         "start"
       ],
-      "env": {
-        "NEO4J_URI": "YOUR_NEO4J_URI",
-        "NEO4J_USERNAME": "YOUR_NEO4J_USERNAME",
-        "NEO4J_PASSWORD": "YOUR_NEO4J_PASSWORD"
-      },
       "disabled": false,
       "alwaysAllow": []
     }
@@ -580,6 +743,35 @@ If you installed CodeGraphContext using `pipx`, use the following configuration
 
 ---
 
+## Token-Optimized Output (GCF)
+
+Reduce tool response tokens by ~62% with the opt-in [GCF](https://gcformat.com) output format:
+
+```bash
+pip install gcf-python
+CGC_OUTPUT_FORMAT=gcf codegraphcontext mcp start
+```
+
+Or add it to your MCP client config:
+
+```json
+{
+  "mcpServers": {
+    "CodeGraphContext": {
+      "command": "codegraphcontext",
+      "args": ["mcp", "start"],
+      "env": {
+        "CGC_OUTPUT_FORMAT": "gcf"
+      }
+    }
+  }
+}
+```
+
+GCF encodes structured data with positional fields (keys declared once, values pipe-delimited). Code graph query results (symbols, relationships, callers, complexity) are exactly the data shape where GCF saves the most. 100% LLM comprehension on all frontier models. Falls back to JSON if `gcf-python` is not installed.
+
+---
+
 ## Natural Language Interaction Examples
 
 Once the server is running, you can interact with it through your AI assistant using plain English. Here are some examples of what you can say:

{codegraphcontext-0.5.1 → codegraphcontext-0.5.3}/README.md

@@ -11,7 +11,7 @@
 - 🇯🇵 [日本語](docs/translations/README.ja.md)
 - 🇪🇸 Español (Soon)
 
-🌍 **Help translate CodeGraphContext to your language by raising an issue & PR on [GitHub Issues](https://github.com/Shashankss1205/CodeGraphContext/issues)!**
+🌍 **Help translate CodeGraphContext to your language by raising an issue & PR on [GitHub Issues](https://github.com/CodeGraphContext/CodeGraphContext/issues)!**
 
 <p align="center">
   <br>
@@ -99,6 +99,43 @@ A powerful **MCP server** and **CLI toolkit** that indexes local code into a gra
 
 ---
 
+## Why CodeGraphContext?
+
+CodeGraphContext is built for the moment when plain text search stops being enough. It turns a repository into a graph of files, symbols, calls, inheritance, imports, and relationships so you can move from "where is this defined?" to "how does this actually connect?" without jumping between tools.
+
+```mermaid
+flowchart LR
+  A[Source code] --> B[Tree-sitter and SCIP indexers]
+  B --> C[Graph database]
+  C --> D[CLI queries]
+  C --> E[MCP server]
+  D --> F[Direct analysis]
+  E --> G[AI assistants with repo context]
+```
+
+### Comparison at a glance
+
+| Approach | Best for | Tradeoff |
+| --- | --- | --- |
+| `grep` / file search | Exact string lookup | Misses relationships and code structure |
+| RAG over code chunks | Natural-language retrieval | Can lose symbol-level precision |
+| CodeGraphContext | Repository-wide reasoning | Requires an indexing step |
+
+### Common use cases
+
+- Trace call chains across files and packages.
+- Understand class hierarchies, imports, and module boundaries.
+- Find dead code, hotspots, and unexpected dependencies.
+- Give an AI assistant the same structural context a maintainer would use.
+
+### FAQ
+
+- **Do I need a remote service?** No. CGC works locally with embedded backends or with external graph databases when you want them.
+- **Is it only for one language?** No. It supports a broad mix of mainstream languages and keeps expanding.
+- **Can I use it with an AI assistant?** Yes. That is one of the main reasons the MCP server exists.
+
+---
+
 ## Project Details
 - **Version:** 0.5.1
 - **Authors:** Shashank Shekhar Singh <shashankshekharsingh1205@gmail.com>
@@ -136,6 +173,49 @@ A powerful **MCP server** and **CLI toolkit** that indexes local code into a gra
 -   **Flexible Database Backend:** FalkorDB Lite (Default), KuzuDB, LadybugDB, FalkorDB Remote, Nornic DB, or Neo4j (all platforms via Docker/native).
 
 
+---
+## Architecture Overview
+
+CodeGraphContext transforms source code into a queryable knowledge graph that can be explored through the CLI or AI assistants via MCP.
+
+```mermaid
+flowchart TD
+    A[Code Repository] --> B[Tree-sitter / SCIP Indexing]
+    B --> C[Knowledge Graph]
+    C --> D[Graph Database]
+    D --> E[CLI Toolkit]
+    D --> F[MCP Server]
+    F --> G[AI Assistant]
+```
+
+### Workflow
+
+1. Source code is parsed using Tree-sitter or SCIP indexers.
+2. Relationships between functions, classes, imports, and calls are extracted.
+3. A knowledge graph is generated and stored in a graph database.
+4. Users can query the graph through the CLI or AI assistants using MCP.
+
+## 🏗️ Architecture & Workflow
+
+CodeGraphContext parses your source code and builds a comprehensive knowledge graph. This graph can be queried directly via the CLI toolkit or exposed to AI assistants through the MCP server.
+
+```mermaid
+flowchart TD
+    A[Code Repository] -->|Parsed by| B[Tree-sitter / SCIP Indexing]
+    B -->|Generates| C[Knowledge Graph]
+    C -->|Stored in| D[(Graph Database)]
+    D -->|Queried via| E[CLI Toolkit]
+    D -->|Served by| F[MCP Server]
+    F -->|Provides context to| G[🤖 AI Assistant]
+
+    classDef default fill:#1f2937,stroke:#8b5cf6,stroke-width:2px,color:#fff;
+    classDef db fill:#0f172a,stroke:#3b82f6,stroke-width:2px,color:#fff;
+    classDef ai fill:#312e81,stroke:#a855f7,stroke-width:2px,color:#fff;
+    
+    D:::db
+    G:::ai
+```
+
 ---
 
 ## Supported Programming Languages
@@ -206,8 +286,8 @@ _If you’re using CodeGraphContext in your project, feel free to open a PR and
 - `rich>=13.7.0`
 - `inquirerpy>=0.3.4`
 - `python-dotenv>=1.0.0`
-- `tree-sitter>=0.21.0` (not installed on Python 3.13)
-- `tree-sitter-language-pack>=0.6.0` (not installed on Python 3.13)
+- `tree-sitter>=0.24.0,<0.26.0`
+- `tree-sitter-language-pack>=1.6,<2.0`
 - `pyyaml`
 - `pathspec>=0.12.1`
 - `falkordb>=1.0,<1.6`
@@ -241,6 +321,87 @@ _If you’re using CodeGraphContext in your project, feel free to open a PR and
     - **KuzuDB:** Cross-platform embedded backend.
     - **Neo4j:** Run `codegraphcontext neo4j setup` to use an external server.
 
+### 🧪 Local Development Setup
+
+If you want to run CodeGraphContext from a local clone and contribute changes, use an editable install:
+
+```bash
+git clone https://github.com/CodeGraphContext/CodeGraphContext.git
+cd CodeGraphContext
+python -m venv .venv
+source .venv/bin/activate
+pip install -e ".[dev]"
+```
+
+Then verify the install and try the local commands:
+
+```bash
+cgc --help
+cgc doctor
+cgc index .
+cgc analyze callers main
+```
+
+For MCP server work, finish the setup with:
+
+```bash
+cgc mcp setup
+cgc mcp start
+```
+
+If you are only installing the published package, `pip install codegraphcontext` is enough. If you are developing parser-heavy features and want the optional parsing extras too, install with `pip install -e ".[dev,parsing]"`.
+
+---
+
+## 🐳 Docker (No Python Required)
+
+Run CodeGraphContext without installing Python — pull the image and start indexing:
+
+### Quick Start
+```bash
+# Pull the latest image
+docker pull codegraphcontext/codegraphcontext:latest
+
+# Index your current directory
+docker run --rm -v "$(pwd):/workspace" codegraphcontext/codegraphcontext cgc index .
+
+# List indexed repos
+docker run --rm -v cgc-data:/home/cgc/.codegraphcontext \
+    codegraphcontext/codegraphcontext cgc list
+
+# Interactive shell
+docker run -it --rm -v "$(pwd):/workspace" \
+    -v cgc-data:/home/cgc/.codegraphcontext \
+    codegraphcontext/codegraphcontext bash
+```
+
+### Docker Compose
+```bash
+# Clone the repo and start with Docker Compose
+git clone https://github.com/CodeGraphContext/CodeGraphContext.git
+cd CodeGraphContext
+
+# Start with embedded DB
+docker compose run --rm cgc index .
+
+# Start with external FalkorDB
+docker compose --profile falkordb up -d
+
+# Start with visualization server
+docker compose --profile viz up -d
+# Open http://localhost:8080
+```
+
+### Available Docker Tags
+| Tag | Description |
+|---|---|
+| `latest` | Latest stable release |
+| `edge` | Latest from `main` branch (may be unstable) |
+| `0.4.19` | Specific version |
+| `0.4` | Latest patch in 0.4.x |
+
+For more advanced Docker usage, including running the MCP Server or connecting external databases, see our [Comprehensive Docker Guide](docs/DOCKER.md).
+
 ---
 
 ## 📋 Prerequisites
@@ -264,7 +425,7 @@ python --version
 ### Step 1: Install CodeGraphContext
 
 ```bash
-pip install codegraphcontext
+pip install codegraphcontext  # Installs CodeGraphContext and required dependencies
 ```
 
 This command installs CodeGraphContext and all required dependencies.
@@ -288,7 +449,7 @@ CodeGraphContext automatically uses an embedded database by default, so no addit
 ### Index a Repository
 
 ```bash
-codegraphcontext index .
+codegraphcontext index .  # Scans current directory and builds code graph
 ```
 
 This scans the current project and creates a searchable code graph.
@@ -316,7 +477,7 @@ Finds potentially unused code in the indexed repository.
 After indexing a repository, run:
 
 ```bash
-codegraphcontext list
+codegraphcontext list  # Shows all indexed repositories in the system
 ```
 
 If the command executes successfully and displays indexed repositories, your setup is complete and CodeGraphContext is ready to use.
@@ -445,11 +606,6 @@ Add the following server configuration to your client's settings file (e.g., VS
         "mcp",
         "start"
       ],
-      "env": {
-        "NEO4J_URI": "YOUR_NEO4J_URI",
-        "NEO4J_USERNAME": "YOUR_NEO4J_USERNAME",
-        "NEO4J_PASSWORD": "YOUR_NEO4J_PASSWORD"
-      },
       "disabled": false,
       "alwaysAllow": []
     }
@@ -489,6 +645,35 @@ If you installed CodeGraphContext using `pipx`, use the following configuration
 
 ---
 
+## Token-Optimized Output (GCF)
+
+Reduce tool response tokens by ~62% with the opt-in [GCF](https://gcformat.com) output format:
+
+```bash
+pip install gcf-python
+CGC_OUTPUT_FORMAT=gcf codegraphcontext mcp start
+```
+
+Or add it to your MCP client config:
+
+```json
+{
+  "mcpServers": {
+    "CodeGraphContext": {
+      "command": "codegraphcontext",
+      "args": ["mcp", "start"],
+      "env": {
+        "CGC_OUTPUT_FORMAT": "gcf"
+      }
+    }
+  }
+}
+```
+
+GCF encodes structured data with positional fields (keys declared once, values pipe-delimited). Code graph query results (symbols, relationships, callers, complexity) are exactly the data shape where GCF saves the most. 100% LLM comprehension on all frontier models. Falls back to JSON if `gcf-python` is not installed.
+
+---
+
 ## Natural Language Interaction Examples
 
 Once the server is running, you can interact with it through your AI assistant using plain English. Here are some examples of what you can say: