mcp-server-sqlseed 0.1.20

MCP server for sqlseed - AI-powered SQLite test data generation

mcp-server-sqlseed

English | 中文

Model Context Protocol (MCP) server for sqlseed — enabling AI assistants to inspect schemas, generate configs, and fill SQLite databases.

Installation

# Basic
pip install mcp-server-sqlseed

# With AI support (includes sqlseed-ai)
pip install mcp-server-sqlseed[ai]

Configuration

Claude Desktop

Add to ~/Library/Application Support/Claude/claude_desktop_config.json (macOS) or equivalent:

{
  "mcpServers": {
    "sqlseed": {
      "command": "mcp-server-sqlseed"
    }
  }
}

Cursor / Other MCP Clients

Use command: mcp-server-sqlseed

MCP Tools

Tool	Description
sqlseed_inspect_schema	Inspect database schema: columns, foreign keys, indexes, sample data, schema_hash. Accepts optional table_name (all tables if omitted).
sqlseed_generate_yaml	AI-driven YAML config generation with self-correction. Requires sqlseed-ai plugin and API key. Supports api_key/base_url/model parameter overrides.
sqlseed_execute_fill	Execute data generation. Accepts optional yaml_config string, count, and enrich flag. Max YAML config size: 256KB.

MCP Resource

Resource	Description
sqlseed://schema/{db_path}/{table_name}	Read-only JSON schema for a specific table

Example Usage

After configuring your MCP client, you can prompt:

"Inspect the schema of app.db, generate a YAML config for the users table, then fill 1000 rows."

The AI assistant will call:

1. sqlseed_inspect_schema → get table structure
2. sqlseed_generate_yaml → generate YAML config (if sqlseed-ai is installed)
3. sqlseed_execute_fill → fill data

AI Integration

When sqlseed-ai is installed and an API key is configured (SQLSEED_AI_API_KEY or OPENAI_API_KEY), the sqlseed_generate_yaml tool uses LLM-driven analysis with self-correction. Without the AI plugin, the tool returns a fallback message.

Requirements

• Python >= 3.10
• sqlseed >= 0.1.0
• mcp >= 1.0

Optional:

• sqlseed-ai (for sqlseed_generate_yaml tool)

License

AGPL-3.0-or-later