dbtree

A tool to visualize database schemas right inside your terminal. It's built for both humans and AI to get complete context of your database architecture.

Supports PostgreSQL, MySQL, ClickHouse, and SQLite.

features

• Visualize table relationships and foreign keys
• Multiple output formats: text or json
• Multiple output shapes: tree, flat, chart
• Shows columns, data types, and constraints (primary keys, foreign keys, unique constraints)
• Handles circular references

installation

Linux/macOS:

curl -fsSL https://vivekn.dev/dbtree/install.sh | bash

Or from source:

git clone https://github.com/viveknathani/dbtree.git
cd dbtree
make build # binary will be created at `./bin/dbtree`

Or using go install:

go install github.com/viveknathani/dbtree/cmd/dbtree@latest

usage

• --conn (required): Database connection URL

  PostgreSQL:

  • Format: postgres://username:password@host:port/database
  • Example: postgres://user:pass@localhost:5432/mydb

  MySQL:

  • Format: mysql://username:password@tcp(host:port)/database
  • Example: mysql://root:password@tcp(localhost:3306)/mydb

  ClickHouse:

  • Format: clickhouse://username:password@host:port/database
  • Example: clickhouse://default:@localhost:9000/mydb

  SQLite:

  • Format: sqlite://path/to/database.db or just path/to/database.db
  • Example: sqlite://./mydb.db or ./mydb.db

• --format (optional): Output format

  • text (default): Human-readable text output
  • json: Structured JSON output

• --shape (optional): Visualization structure

  • tree (default): Hierarchical tree showing foreign key relationships
  • flat: Simple list of all tables
  • chart: ASCII diagram of the schema (works with text format only, also see more notes below for this shape)

• --help: Display help information

examples

Tree View (Text)

Shows tables in a hierarchical structure based on foreign key relationships:

dbtree --conn "postgres://user:pass@localhost:5432/mydb" --format text --shape tree

Output:

testdb
├── categories
│   ├── id ("integer") PRIMARY KEY
│   ├── name ("varchar(100)")
│   └── description ("text")
│   └── products
│       ├── id ("integer") PRIMARY KEY
│       ├── name ("varchar(200)")
│       ├── price ("numeric(10,2)")
│       ├── category_id ("integer") → categories.id
│       ├── stock ("integer")
│       └── created_at ("timestamp")
│       ├── order_items
│       │   ├── id ("integer") PRIMARY KEY
│       │   ├── order_id ("integer") → orders.id
│       │   ├── product_id ("integer") → products.id
│       │   ├── quantity ("integer")
│       │   └── price ("numeric(10,2)")
│       └── reviews
│           ├── id ("integer") PRIMARY KEY
│           ├── product_id ("integer") → products.id
│           ├── user_id ("integer") → users.id
│           ├── rating ("integer")
│           ├── comment ("text")
│           └── created_at ("timestamp")
└── users
    ├── id ("integer") PRIMARY KEY
    ├── username ("varchar(50)") UNIQUE
    ├── email ("varchar(100)") UNIQUE
    └── created_at ("timestamp")
    ├── addresses
    │   ├── id ("integer") PRIMARY KEY
    │   ├── user_id ("integer") → users.id
    │   ├── street ("varchar(200)")
    │   ├── city ("varchar(100)")
    │   ├── country ("varchar(100)")
    │   └── zip_code ("varchar(20)")
    ├── orders
    │   ├── id ("integer") PRIMARY KEY
    │   ├── user_id ("integer") → users.id
    │   ├── total_amount ("numeric(10,2)")
    │   ├── status ("varchar(50)")
    │   └── created_at ("timestamp")
    │   ├── order_items (see above)
    │   └── payments
    │       ├── id ("integer") PRIMARY KEY
    │       ├── order_id ("integer") → orders.id
    │       ├── payment_method ("varchar(50)")
    │       ├── amount ("numeric(10,2)")
    │       ├── status ("varchar(50)")
    │       └── created_at ("timestamp")
    └── reviews (see above)

Flat List (Text)

Lists all tables alphabetically with their columns:

dbtree --conn "postgres://user:pass@localhost:5432/mydb" --format text --shape flat

Output:

Database: testdb
Tables: 8

addresses
  - id (integer) PRIMARY KEY
  - user_id (integer) → users.id
  - street (varchar(200))
  - city (varchar(100))
  - country (varchar(100))
  - zip_code (varchar(20))

categories
  - id (integer) PRIMARY KEY
  - name (varchar(100))
  - description (text)

order_items
  - id (integer) PRIMARY KEY
  - order_id (integer) → orders.id
  - product_id (integer) → products.id
  - quantity (integer)
  - price (numeric(10,2))

orders
  - id (integer) PRIMARY KEY
  - user_id (integer) → users.id
  - total_amount (numeric(10,2))
  - status (varchar(50))
  - created_at (timestamp)

payments
  - id (integer) PRIMARY KEY
  - order_id (integer) → orders.id
  - payment_method (varchar(50))
  - amount (numeric(10,2))
  - status (varchar(50))
  - created_at (timestamp)

products
  - id (integer) PRIMARY KEY
  - name (varchar(200))
  - price (numeric(10,2))
  - category_id (integer) → categories.id
  - stock (integer)
  - created_at (timestamp)

reviews
  - id (integer) PRIMARY KEY
  - product_id (integer) → products.id
  - user_id (integer) → users.id
  - rating (integer)
  - comment (text)
  - created_at (timestamp)

users
  - id (integer) PRIMARY KEY
  - username (varchar(50)) UNIQUE
  - email (varchar(100)) UNIQUE
  - created_at (timestamp)

Chart View (ASCII Diagram)

Renders an ASCII chart diagram of the schema:

dbtree --conn "postgres://user:pass@localhost:5432/mydb" --format text --shape chart

As of right now, this chart view looks ugly in a terminal and is probably not the best for an AI agent. However, when redirected to a text file, it looks decent enough for humans. I am open to feedback on how this can be improved.

JSON Output

Export schema information as structured JSON:

dbtree --conn "postgres://user:pass@localhost:5432/mydb" --format json --shape tree

Output:

{
  "database": "mydb",
  "tables": [
    {
      "name": "users",
      "columns": [
        {
          "name": "id",
          "type": "integer",
          "constraint": "PRIMARY KEY"
        },
        {
          "name": "email",
          "type": "character varying",
          "constraint": "UNIQUE"
        },
        {
          "name": "name",
          "type": "character varying"
        }
      ],
      "children": [
        {
          "name": "posts",
          "columns": [
            {
              "name": "id",
              "type": "integer",
              "constraint": "PRIMARY KEY"
            },
            {
              "name": "user_id",
              "type": "integer",
              "reference": "users.id"
            }
          ]
        }
      ]
    }
  ]
}

using with different databases

MySQL

dbtree --conn "mysql://root:password@tcp(localhost:3306)/mydb" --shape tree

ClickHouse

dbtree --conn "clickhouse://default:@localhost:9000/mydb" --shape tree

Note: ClickHouse does not enforce foreign keys, so only primary keys and table/column information will be shown.

SQLite

dbtree --conn "./mydb.db" --shape tree
# or
dbtree --conn "sqlite://./mydb.db" --shape tree

hacking

I am open to PRs for improving this project.

You will need:

1. make
2. Go (>=v1.25)
3. Docker and Docker Compose (for running tests)

Some handy commands:

# Build the project
make build

# Run tests (starts containers, runs tests, stops containers)
make test

# For local development (keeps containers running)
make test-local
# ... do your testing ...
make test-down  # when done

license

None. Do whatever you want. Have fun and happy hacking!

Built by @viveknathani