Note: This is a personal fork of pondpilot/flowscope maintained by @cooker-code with additional customizations (Trellis AI workflow integration, audit review page, derived-subquery JOIN fix, Cursor IDE platform support). Upstream is tracked via the upstream git remote and synced periodically; see Syncing with upstream below.

FlowScope includes a full web application at flowscope.pondpilot.io for interactive, multi-file SQL lineage analysis.

Under the hood, it is a privacy-first SQL lineage engine that runs entirely in the browser. Built with Rust and WebAssembly, it analyzes SQL queries to produce lineage graphs that describe how tables, CTEs, and columns flow through transformations.

The engine is designed for embedding into web apps, browser extensions, and developer tools that need instant lineage analysis without sending SQL to a server.

The easiest way to use FlowScope is through the hosted web app — no installation required:

flowscope.pondpilot.io

Features:

• Drag and drop SQL files or paste queries directly
• Interactive lineage graph with table and column-level views
• Multi-file project support with schema DDL
• dbt/Jinja template preprocessing for dbt models
• Export to Mermaid, JSON, CSV, Excel, or HTML reports
• Librarian — AI chat panel that answers questions about your data based on lineage analysis and uploaded PDF docs
• All processing happens in your browser — your SQL never leaves your machine

For scripting and CI/CD integration, install the CLI:

# Core CLI (analysis, linting, fixing, exports)
cargo install flowscope-cli

# With the bundled local web server (see "Serve Mode" below)
cargo install flowscope-cli --features serve

Basic usage:

# Analyze a SQL file
flowscope query.sql

# Analyze with a specific dialect
flowscope -d snowflake etl/*.sql

# Generate a Mermaid diagram
flowscope -f mermaid -v column query.sql > lineage.mmd

# Export to Excel with schema awareness
flowscope -s schema.sql -f xlsx -o report.xlsx queries/*.sql

# Pipe from stdin
cat query.sql | flowscope -d postgres

Output formats: table (default), json, mermaid, html, sql, csv, xlsx, duckdb

FlowScope includes a SQL linter with 72 rules covering aliasing, layout, conventions, structure, and more:

# Lint SQL files
flowscope --lint queries/*.sql

# Lint and auto-fix
flowscope --lint --fix queries/*.sql

# JSON output for CI integration
flowscope --lint -f json queries/*.sql

See CLI documentation for all lint options and rule configuration.

Run FlowScope as a local HTTP server with the full web UI embedded in a single binary:

# Start server watching SQL directories
flowscope --serve --watch ./sql

# With database schema and custom port
flowscope --serve --watch ./models -d postgres --metadata-url postgres://user@localhost/db --port 8080

# Open browser automatically
flowscope --serve --watch ./sql --open

The serve mode watches directories for .sql file changes and provides the same interactive experience as the hosted web app, with all processing happening locally. Requires building with the serve feature.

See CLI documentation for all options.

• Client-side analysis with zero data egress
• Multi-dialect coverage (PostgreSQL, Snowflake, BigQuery, DuckDB, Redshift, and more)
• dbt and Jinja templating support with built-in macro stubs (ref(), source(), var())
• Table and column lineage with schema-aware wildcard expansion
• SQL linting with 72 rules across 9 categories (aliasing, layout, convention, structure, and more)
• Auto-fix engine with safe and unsafe fix modes
• Structured diagnostics with spans for precise highlighting
• Completion API for SQL authoring workflows
• TypeScript API and optional React visualization components
• Librarian AI chat panel for natural-language Q&A over SQL lineage and uploaded PDF documentation (OpenAI, Anthropic, or custom endpoints)

• app/ — the hosted web application at flowscope.pondpilot.io
• crates/ — Rust engine, WASM bindings, and CLI
• packages/ — TypeScript API and React visualization components

Install the core package:

npm install @pondpilot/flowscope-core

Analyze a query:

import { initWasm, analyzeSql } from '@pondpilot/flowscope-core';

await initWasm();

const result = await analyzeSql({
  sql: 'SELECT * FROM analytics.orders',
  dialect: 'postgres',
});

console.log(result.statements[0]);

Use the completion API to provide SQL authoring hints at a cursor position. See docs/guides/schema-metadata.md for schema setup details.

import {
  charOffsetToByteOffset,
  completionItems,
  initWasm,
} from '@pondpilot/flowscope-core';

await initWasm();

const sql = 'SELECT * FROM analytics.';
const cursorOffset = charOffsetToByteOffset(sql, sql.length);

const result = await completionItems({
  sql,
  dialect: 'postgres',
  cursorOffset,
  schema: {
    defaultSchema: 'analytics',
    tables: [{ name: 'orders', columns: [{ name: 'order_id' }, { name: 'total' }] }],
  },
});

console.log(result.items.slice(0, 5));

For interactive lineage graphs, add the React package and render the LineageExplorer component. See docs/guides/quickstart.md for a full walkthrough.

npm install @pondpilot/flowscope-react

• docs/README.md — documentation map and reference index
• docs/librarian.md — Librarian AI chat panel user guide
• docs/guides/quickstart.md — TypeScript quickstart guide
• docs/guides/schema-metadata.md — schema metadata setup
• docs/dialect-coverage.md — dialect and statement coverage
• crates/flowscope-cli/README.md — CLI usage and examples
• docs/linter-architecture.md — linter engine design and rule families
• docs/workspace-structure.md — monorepo layout and build entry points

FlowScope uses just for common tasks. Run just build, just test, or just dev, and see docs/workspace-structure.md for the full command list.

See CONTRIBUTING.md for setup, testing expectations, and contribution guidelines.

The core engine and packages are released under Apache-2.0. See LICENSE for details. The app/ directory uses the O'Saasy License; see app/LICENSE.

This fork tracks pondpilot/flowscope via the upstream git remote. To pull in upstream changes:

git fetch upstream
git log --oneline master..upstream/master   # preview what's new
git merge upstream/master                    # or: git rebase upstream/master
git push origin master

If you set up a one-shot alias:

git config alias.sync-upstream '!git fetch upstream && git merge upstream/master && git push origin master'
git sync-upstream

The upstream remote has its push URL intentionally disabled to prevent accidental pushes to pondpilot/flowscope. Restore it explicitly if you ever need to open a PR:

git remote set-url --push upstream https://github.com/pondpilot/flowscope.git

Part of the PondPilot project.