About seedweaver-mcp
SeedWeaver is an MCP server that generates relationally coherent test data from SQL DDL or a JSON schema. It is for developers preparing database fixtures, and accounts for foreign keys and uniqueness constraints while producing SQL, JSON, or CSV.
Key features
- 01Parses SQL DDL or JSON schemas
- 02Generates data in foreign-key dependency order
- 03Respects primary-key, unique, and not-null constraints
- 04Creates common values such as names, emails, and dates
- 05Outputs SQL, JSON, or CSV
Use cases
- 01Seed databases with related tables
- 02Prepare orders and users with resolvable foreign keys
- 03Control table row counts and enums through JSON schemas
- 04Generate SQL or JSON data for tests
MCP support
As an MCP server, it exposes schema analysis and test-data generation tools to MCP clients.
Configuration
Local server config is the same for every client. Copy it into your client’s MCP config file.
Tools
2Saved tool list · updated 2w ago
| Parameter | Type | Required | Description |
|---|---|---|---|
| schema | string | Yes | The schema to analyze: either raw SQL DDL (CREATE TABLE ...) or a JSON schema object with a 'tables' array. |
Details
From the project READMESeedWeaver 🌱
Generate realistic, referentially-coherent test data from your real database schema — straight from Claude, Cursor, or any MCP client.
Paste your CREATE TABLE SQL (or a JSON schema) and SeedWeaver generates seed data where foreign keys actually resolve to real primary keys, unique constraints are respected, and values are realistic (names, emails, dates, decimals). Unlike generic mock-data tools that spit out statistically random junk, SeedWeaver understands the relationships in your schema.
Ask your AI: "Generate 50 users and 200 orders for this schema" — and get INSERTs you can run immediately, with every
order.user_idpointing at a user that exists.
Why SeedWeaver
Most "fake data" tools (Faker wrappers, random generators) give you isolated rows with no awareness of your schema. The moment you have foreign keys, you're back to hand-wiring relationships. SeedWeaver:
- Resolves foreign keys — referenced tables are generated first (topological ordering), and FK columns point at real generated keys.
- Respects constraints — unique columns stay unique, primary keys are unique,
NOT NULLis honored. - Reads your real schema — paste Postgres/MySQL
CREATE TABLEDDL directly, no manual config. - Realistic values — names, emails, addresses, companies, dates, decimals, enums.
- Outputs what you need — SQL
INSERTstatements, JSON, or CSV.
Install
npx -y seedweaver-mcp
Add to your MCP client config (Claude Desktop / Cursor / Windsurf):
{
"mcpServers": {
"seedweaver": {
"command": "npx",
"args": ["-y", "seedweaver-mcp"]
}
}
}
Tools
| Tool | What it does |
|---|---|
analyze_schema | Parse a schema (SQL DDL or JSON) and report tables, columns, relationships, and generation order. Run this first to confirm SeedWeaver reads your schema correctly. |
generate_seed_data | Generate coherent test data. Returns SQL INSERTs (default), JSON, or CSV. |
Examples
From SQL DDL:
Generate 20 rows of test data for:
CREATE TABLE users (
id UUID PRIMARY KEY,
email VARCHAR(255) UNIQUE NOT NULL,
full_name VARCHAR(100),
created_at TIMESTAMP
);
CREATE TABLE orders (
id SERIAL PRIMARY KEY,
user_id UUID NOT NULL REFERENCES users(id),
total DECIMAL(10,2),
status VARCHAR(20)
);
Every generated orders.user_id will be a real users.id.
From a JSON schema (gives you fine control — enums, ranges, per-table row counts):
{
"tables": [
{
"name": "users",
"rows": 20,
"columns": [
{ "name": "id", "type": "uuid", "primaryKey": true },
{ "name": "email", "type": "email", "unique": true },
{ "name": "name", "type": "fullName" }
]
},
{
"name": "orders",
"rows": 80,
"columns": [
{ "name": "id", "type": "serial", "primaryKey": true },
{ "name": "user_id", "type": "fk", "references": "users.id" },
{ "name": "total", "type": "decimal", "min": 5, "max": 500 },
{ "name": "status", "type": "enum", "values": ["pending", "paid", "shipped"] }
]
}
]
}
Supported column types: uuid, serial, int, decimal, boolean, email, fullName, firstName, lastName, username, phone, address, city, country, company, url, word, sentence, paragraph, date, datetime, enum, fk.
Free vs. Pro
The free tier is fully functional for small schemas. Pro removes the limits and adds the features you need for real projects and CI.
| Free | Pro | |
|---|---|---|
| Tables per schema | 2 | Unlimited |
| Rows per table | 50 | Unlimited |
| Output formats | SQL, JSON | + CSV |
| Deterministic seeds (reproducible data) | — | ✓ |
| Custom locales | — | ✓ |
| Use in CI / automation | — | ✓ |
→ Get SeedWeaver Pro — $19/mo
Activate by setting your license key:
{
"mcpServers": {
"seedweaver": {
"command": "npx",
"args": ["-y", "seedweaver-mcp"],
"env": { "SEEDWEAVER_LICENSE": "your-key-here" }
}
}
}
License
The SeedWeaver MCP server is MIT licensed and free to run. Pro features are unlocked with a paid license key. Built with the Model Context Protocol.
Comments
New to MCP?
From what the protocol is to installing your first server, these answers cover most first questions.
An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.