Quickstart Guide

Create your first project

After logging in, navigate to Settings → Schemas and click New Schema. Give it a name that describes your project (e.g., "Core Banking" or "Insurance Platform").

A schema is a namespace that isolates one project's assets from another. Each schema has its own index, its own search scope, and its own configuration.

Connect your source code

Codegraphs supports several ways to ingest source code:

• Git Clone — provide a repository URL and access token. Codegraphs clones and keeps the working copy in sync.
• Local Directory — point to a directory on the server (for on-premise deployments).
• ZIP Upload — upload a ZIP archive through the UI.
• SharePoint — connect a Microsoft 365 site/drive via OAuth.
• WebDAV — connect a network share or on-prem document server.

For this quickstart, we'll use Git Clone:

# In the Codegraphs Manager UI:
Schema → Sources → Add Source → Git
URL: https://github.com/your-org/your-repo.git
Branch: main
# Codegraphs auto-detects languages and extensions on first index.

Start the indexing pipeline

From the Schema page, click Index Now. The pipeline runs through several stages:

Reading files...       ████████████ 100%
Classifying assets...  ████████████ 100%  (48,708 files, 43 extensions)
Parsing sources...     ████████████ 100%  (COBOL 2.3M LOC, JCL 1,240 jobs)
Extracting relations...████████████ 100%  (12,450 CALLs, 3,200 COPY refs)
Building index...      ████████████ 100%  (ready)

A 100K-line codebase typically indexes in under 30 seconds. Multi-million-line systems may take several minutes on the first run. Subsequent runs are incremental — only changed files are re-indexed.

Search across all assets

Once indexing completes, open the Search page. Try these queries:

CUSTOMER                          # Find everything mentioning "CUSTOMER"
ext:cbl CUSTOMER                  # Limit to COBOL files only
dir:batch/                        # Search within the batch/ directory
file:ACCT                         # Filenames containing "ACCT"
"UPDATE" "TABLE"                  # Exact phrase AND search

Search returns results grouped by asset type — programs, copybooks, JCL jobs, SQL, documents — with highlighted matches and direct source viewing.

Navigate relationships

Click any program name in search results to open the Graph View. You'll see:

• Callees — programs this program calls
• Callers — programs that call this program
• COPY includes — copybooks this program uses
• File I/O — datasets read, written, updated, or deleted
• SQL tables — database tables accessed
• JCL jobs — batch jobs that execute this program

Click any node to expand it. The graph grows with your exploration. Use the depth slider to control how many levels of transitive relationships to reveal.