db2md

Generate Markdown documentation from any database — ER diagrams, table details, and AI-powered overviews.

Features

• Multi-database support — PostgreSQL, MySQL, MSSQL, SQLite, MongoDB
• ER diagram — Mermaid erDiagram showing relationships between tables/collections
• Table details — Column types, nullability, primary/foreign keys, indexes
• AI overview — Optional natural-language summary via AWS Bedrock (Claude)
• Interactive CLI — Guided prompts if no flags are provided

Install

npm install -g @nikhil-cephei/db2md

Or use directly without installing:

npx @nikhil-cephei/db2md

Usage

db2md --type <db-type> --url "<connection-string>"

Or run interactively (prompts for all connection details):

db2md

Connection Options

| Flag | Description | |------|-------------| | --type <type> | Database type: postgres, mysql, mssql, sqlite, mongodb | | --url <url> | Full connection string (takes precedence over individual flags) | | --host <host> | Database host (default: localhost) | | --port <port> | Database port (default: varies by type) | | --db <name> | Database name | | --user <user> | Username | | --password <pass> | Password |

Output Options

| Flag | Description | |------|-------------| | --output <dir> | Output directory (default: .) | | --filename <name> | Output filename (default: db-schema.md) |

AI Options

| Flag | Description | |------|-------------| | --context <text> | Project context for a more accurate AI description | | --no-ai | Skip AI analysis even if AWS credentials are set |

Examples

# SQLite
db2md --type sqlite --url ./mydb.db

# PostgreSQL via connection string
db2md --type postgres --url "postgresql://user:pass@localhost:5432/mydb"

# MySQL with individual flags
db2md --type mysql --host localhost --db myapp --user root

# MSSQL
db2md --type mssql --url "mssql://user:pass@localhost:1433/mydb"

# MongoDB
db2md --type mongodb --url "mongodb://localhost:27017/mydb"

# Custom output directory with AI context
db2md --type postgres --url "..." --output ./docs --context "Healthcare SaaS"

# Disable AI even if credentials are configured
db2md --type postgres --url "..." --no-ai

AI Overview (AWS Bedrock)

When AWS credentials are available, db2md automatically calls AWS Bedrock (Claude) to generate a plain-English description of your schema. Use --context to provide project background for a more accurate result.

Configure credentials

db2md config set    # store credentials encrypted
db2md config show   # view stored credentials
db2md config clear  # remove all stored credentials

Credentials are stored encrypted on disk. Alternatively, set environment variables directly:

| Variable | Description | |----------|-------------| | AWS_ACCESS_KEY_ID | AWS access key | | AWS_SECRET_ACCESS_KEY | AWS secret key | | AWS_REGION | AWS region (default: us-east-1) | | BEDROCK_MODEL_ID | Claude model ID (default: anthropic.claude-3-5-sonnet-20241022-v2:0) |

Output

The generated Markdown file includes:

1. AI Overview (if AWS credentials are available) — high-level description of the database purpose and key features
2. ER Diagram — Mermaid diagram (also saved as a .mmd file alongside the Markdown)
3. Table/Collection Details — columns, types, constraints, indexes

Requirements

• Node.js >= 18

License

MIT — Nikhil Makwana