flowmo

A lightning-fast, zero-infrastructure SQL engine for AI-assisted OutSystems prototyping.

Flowmo gives AI agents a local offline environment to generate and validate Advanced SQL queries in milliseconds — acting as a safe sandbox before anything touches an OutSystems environment.

Powered by PGLite (WASM PostgreSQL) — no containers, no servers, no cloud dependencies.

Quick Start

Scaffold a new project with create-flowmo, then install flowmo inside it:

npx create-flowmo

You'll be prompted for a project name, target platform, and app type. Then:

cd my-prototype
npm install

The generated project already has flowmo listed as a dependency and the database/ folder pre-configured.

Commands

flowmo db:setup

Reads database/schema.sql, drops the existing schema, and provisions the local database from scratch.

npx flowmo db:setup

Run this any time you change your schema.

flowmo db:seed [file …]

Inserts seed data into the local database. Accepts an optional list of seed files to run in order.

# Auto-discover: loads database/seeds/ directory (alphabetical) or falls back to database/seeds.sql
npx flowmo db:seed

# Explicit list — executed in the order given
npx flowmo db:seed database/seeds/01_users.sql database/seeds/02_products.sql

Seed file resolution (no args):

1. database/seeds/ directory exists → all .sql files, alphabetical order
2. database/seeds.sql → single file fallback

Prefix files with numbers (01_, 02_) to control load order when using the directory.

flowmo db:reset [--seed [file …]]

Drops and recreates the schema (equivalent to db:setup), then optionally seeds.

# Recreate schema only
npx flowmo db:reset

# Recreate schema + auto-discover seeds
npx flowmo db:reset --seed

# Recreate schema + explicit seed list
npx flowmo db:reset --seed database/seeds/01_users.sql database/seeds/02_products.sql

The --seed flag follows the same resolution rules as db:seed — files after --seed are used as-is; no files after --seed triggers auto-discovery.

flowmo db:query <file|sql> [params-json]

Executes a .sql or .advance.sql file against the local database and prints results as an ASCII table. Alternatively, pass an inline SQL string directly — no file needed.

Inline SQL:

npx flowmo db:query "SELECT * FROM users"
npx flowmo db:query "SELECT COUNT(*) FROM orders WHERE is_active = 1"

Inline mode is param-free. For parameterised queries, use a file.

Standard SQL:

npx flowmo db:query database/queries/get_users.sql

OutSystems Advanced SQL (.advance.sql) with parameters:

npx flowmo db:query database/queries/get_user.advance.sql '{"UserId": 1, "Status": true}'

The Advanced SQL parser handles OutSystems syntax automatically:

| OutSystems syntax | Translated to | |---|---| | {Entity}.[Attribute] | entity.attribute | | {Entity} | entity | | @ParamName | $1, $2, … |

Example .advance.sql:

SELECT {Users}.[Name], {Users}.[Email]
FROM {Users}
WHERE {Users}.[Id] = @UserId AND {Users}.[IsActive] = @Status

Output:

┌─────────┬──────────────────┐
│ Name    │ Email            │
├─────────┼──────────────────┤
│ Izam B. │ [email protected] │
└─────────┴──────────────────┘
(1 row)

Project Structure

A scaffolded Flowmo project looks like this:

my-prototype/
├── database/
│   ├── schema.sql          # DDL — your single source of truth
│   ├── seeds.sql           # Dummy data (single file)
│   ├── seeds/              # OR: split seeds by table (01_users.sql, 02_products.sql …)
│   └── queries/            # .sql and .advance.sql files
├── logic/                  # .flowchart.md server action flows
├── screens/                # .visual.html UI prototypes
├── theme/                  # OutSystems UI CSS
└── .flowmo/                # Local database (auto-generated, gitignored)
    └── database/

The Agentic Bridge (ODC to Flowmo)

Flowmo is built for vibe coding. Because ODC restricts direct database access, agents use local Flowmo projects to iterate safely:

1. AI Generation: Ask your AI assistant to generate a PostgreSQL CREATE TABLE and some seed data based on your ODC data model.
2. Instant Validation: The agent runs npx flowmo db:setup and db:seed to provision the local PGLite database.
3. Query Verification: The agent then executes npx flowmo db:query against .advance.sql files to double-check their logic.
4. Platform Sync: Once the "vibe" is correct locally, you can confidently recreate the entities and queries in OutSystems Service Studio.

The Flowmo Agent Skills (scaffolded by create-flowmo) ensure your AI assistant follows OutSystems performance and syntax rules throughout this process.

VS Code Ecosystem

Flowmo CLI works alongside the Flowmo VS Code extensions:

• Flowmo Visual Inspector — Live layer panel and element inspector for .visual.html screens.
• Flowmo Flowchart Editor — Drag-and-drop node editor for .flowchart.md server action flows.
• PGlite Explorer (third-party) — Browse your .flowmo/database tables and run queries directly in VS Code.

Install the Visual Inspector and Flowchart Editor together with the Flowmo Extension Pack.

Links and Support

• Web: flowmo.lol
• Issues: GitHub Issues
• Email: [email protected]
• Support model: best-effort, no response-time guarantee

License

MIT