SQLite

A SQLite database MCP service provided by Hawa Code (@dahawa/hawa-sqlite-mcp), based on the MCP (Model Context Protocol) protocol over stdio transport, giving AI the ability to query table schemas and execute SQL.

Recommended for use with Hawa Code Code Mode.

---

Quick Start

The package is published to npm and can be run directly via npx — no local build required.

Option 1: Specify the database path via command-line arguments

{
  "mcpServers": {
    "sqlite": {
      "command": "npx",
      "args": [
        "-y",
        "@dahawa/hawa-sqlite-mcp",
        "/absolute/path/to/data.sqlite"
      ]
    }
  }
}

Option 2: Specify the database path via environment variable

{
  "mcpServers": {
    "sqlite": {
      "command": "npx",
      "args": ["-y", "@dahawa/hawa-sqlite-mcp"],
      "env": {
        "SQLITE_DB_PATH": "/absolute/path/to/data.sqlite"
      }
    }
  }
}

Database path priority: SQLITE_DB_PATH environment variable > command-line argument in args > default ./data.sqlite (relative to the MCP client’s working directory). If the client cannot find npx/node, use the absolute path returned by which node instead.

---

Tools Overview

Tool	Description
list_tables	List all tables/views, optionally including CREATE DDL and column structure
execute_ddl	Execute DDL statements (CREATE / ALTER / DROP, supports batch execution of multiple statements)
execute_sql	Execute SQL (SELECT / INSERT / UPDATE / DELETE), supports ? parameter binding

---

list_tables

Lists all tables and views in the SQLite database, optionally including the CREATE DDL and column structure for each table.

Input

Field	Type	Required	Default	Description
includeSchema	boolean	No	true	Whether to include the CREATE DDL statement and column structure for each table

Output

Field	Type	Description
database	string	Database file path
tables	array<TableInfo>	List of tables/views, sorted by name
error	string (optional)	Error message on failure; tables is [] on failure

TableInfo

Field	Type	Description
name	string	Table/view name
type	string	Object type: "table" or "view"
sql	string | null	CREATE DDL statement; null when includeSchema=false
columns	array<ColumnInfo> | null	Column structure; null when includeSchema=false or for non-table objects

ColumnInfo

Field	Type	Description
name	string	Column name
type	string	Declared column type (e.g. INTEGER, TEXT)
notnull	boolean	Whether the column has a NOT NULL constraint
defaultValue	string | null	Default value SQL expression
primaryKey	boolean	Whether the column is a primary key column

Example

Request:

{ "includeSchema": true }

Successful response (excerpt):

{
  "database": "./data.sqlite",
  "tables": [
    {
      "name": "users",
      "type": "table",
      "sql": "CREATE TABLE users (id INTEGER PRIMARY KEY, name TEXT NOT NULL)",
      "columns": [
        {
          "name": "id",
          "type": "INTEGER",
          "notnull": false,
          "defaultValue": null,
          "primaryKey": true
        }
      ]
    }
  ]
}

Failed response (isError: true):

{
  "database": "./data.sqlite",
  "tables": [],
  "error": "<error message>"
}

---

execute_ddl

Executes DDL statements (CREATE / ALTER / DROP TABLE, CREATE INDEX, etc.), supporting batch execution of multiple statements.

Input

Field	Type	Required	Description
sql	string	Yes	The DDL statement(s) to execute; multiple statements can be submitted at once separated by ;

Output

Field	Type	Description
success	boolean	Whether execution succeeded
changes	number	Number of affected rows (always 0 for DDL)
error	string (optional)	Error message on failure

Example

Request:

{ "sql": "CREATE INDEX IF NOT EXISTS idx_users_name ON users(name)" }

Successful response:

{ "success": true, "changes": 0 }

Failed response (isError: true):

{
  "success": false,
  "changes": 0,
  "error": "incomplete input"
}

---

execute_sql

Executes arbitrary SQL statements (SELECT / INSERT / UPDATE / DELETE, etc.), supporting parameter binding.

Input

Field	Type	Required	Description
sql	string	Yes	The SQL statement to execute, using ? placeholders for parameter binding
params	array<string | number | boolean | null>	No	Parameter values bound to the SQL placeholders, in order

Output

Field	Type	Description
columns	array<string>	List of column names
rows	array<object>	Result rows, each row is a “column name → value” object
rowCount	number	Number of result rows (number of affected rows for write operations)
error	string (optional)	Error message on failure; columns/rows are [] and rowCount is 0 on failure

Behavior

• Query statements (SELECT, etc.): columns contains the result set column names, rows contains the query results, rowCount = rows.length
• Write statements (INSERT / UPDATE / DELETE, etc.): columns is always ["changes", "lastInsertRowid"], rows is a single row [{ changes, lastInsertRowid }], rowCount = changes

Example

SELECT

Request:

{ "sql": "SELECT * FROM users WHERE age > ?", "params": [18] }

Response:

{
  "columns": ["id", "name", "age"],
  "rows": [
    { "id": 1, "name": "Alice", "age": 30 }
  ],
  "rowCount": 1
}

INSERT

Request:

{ "sql": "INSERT INTO users (name, age) VALUES (?, ?)", "params": ["Alice", 30] }

Response:

{
  "columns": ["changes", "lastInsertRowid"],
  "rows": [
    { "changes": 1, "lastInsertRowid": 1 }
  ],
  "rowCount": 1
}

Execution failure (isError: true)

{
  "columns": [],
  "rows": [],
  "rowCount": 0,
  "error": "NOT NULL constraint failed: users.name"
}