SQLFlow Cloud is Gudu SQLFlow delivered as an online SQL lineage tool: paste SQL, upload files, or connect a database in your browser and get an interactive, column-level data lineage diagram. There is nothing to install, a free tier to start with, and a premium plan at $49.99/month when you need larger inputs, API access, and team features.
Skip the reading: open SQLFlow Cloud, paste any SQL query, and watch the column-level lineage appear. That single round trip tells you more than any feature list.
What this online SQL lineage tool gives you in the browser
SQLFlow Cloud is the SaaS edition of Gudu SQLFlow, the SQL data lineage tool, running at sqlflow.gudusoft.com. It is the full analysis engine, not a demo: the same dialect-specific parsers, the same column-level resolution, the same data-flow analyzer that powers the enterprise deployments. In the browser you can:
- Trace column-level lineage: for every output column, see exactly which source columns feed it and through which functions, casts, subqueries, joins, and set operators.
- Toggle indirect (impact) lineage: separately show columns that shape a result through
WHERE,JOIN, andGROUP BYclauses without landing in the output. - Generate ER diagrams from DDL: SQLFlow infers primary/foreign-key relationships from your
CREATE TABLEstatements and draws the entity-relationship diagram. - Explore stored procedure call graphs: an interactive map of which Oracle PL/SQL or SQL Server T-SQL procedures invoke which, with lineage traced through parameters, temp tables, and dynamic SQL.
- Ask the AI lineage query: natural-language questions over the analyzed graph, with every referenced table and column validated against the graph before it is shown.
- Export everything: JSON, CSV, or PNG downloads, plus a REST API on premium plans.
Three ways to get SQL in
Paste it. Drop a query, a view definition, or a whole script into the editor, pick one of 39 dialects (Snowflake, BigQuery, Oracle, SQL Server, PostgreSQL, Teradata, Databricks, Spark SQL, and 31 more), and analyze. This is the fastest path and the one most people start with.
Upload files. Point SQLFlow Cloud at SQL files or a dbt manifest and it analyzes the whole set together, resolving references between scripts — a view defined in one file feeding an insert in another shows up as one connected graph.
Connect a source. Pull metadata from a live database over JDBC, or feed in Snowflake query history or Redshift query logs, so lineage reflects what actually runs in your warehouse rather than what someone remembered to document. In all cases SQLFlow performs static analysis of SQL code and schema metadata only; it never reads the rows in your tables.
Column-level lineage, including the indirect kind
Consider a statement any warehouse has a thousand of:
INSERT INTO monthly_revenue (month, total)
SELECT DATE_TRUNC('month', o.order_date) AS month,
SUM(o.amount) AS total
FROM orders o
JOIN customers c ON c.id = o.customer_id
WHERE c.region = 'EMEA'
GROUP BY 1;SQLFlow Cloud shows that monthly_revenue.total is fed directly by orders.amount through SUM(), and monthly_revenue.month by orders.order_date through DATE_TRUNC(). It also shows the columns that never appear in the output but decide what it contains: customers.region filters the rows, c.id and o.customer_id join them, order_date groups them. SQLFlow models those as indirect lineage, a distinct relationship type you can toggle on and off in the diagram. Most competing tools do not make this distinction, yet it is exactly what impact analysis needs: changing customers.region silently changes monthly_revenue.total.
Resolution is genuinely semantic, not pattern matching. The engine expands SELECT *, follows references through CTEs, subqueries, and views, and understands each dialect’s own syntax because it ships a dedicated parser per dialect rather than one generic ANSI grammar. Underneath is the General SQL Parser, developed commercially since the mid-2000s and validated against roughly 13,600 per-dialect SQL test fixtures.
The hard SQL: procedures, dynamic SQL, dbt
Online SQL tools tend to fall over on exactly the code you most need lineage for. SQLFlow Cloud handles the hard cases because the engine was built for them:
- Stored procedures: dedicated procedural parsers for Oracle PL/SQL and SQL Server T-SQL trace lineage through procedure parameters and temp tables, and render an interactive call graph of procedure-to-procedure invocations.
- Dynamic SQL: SQL strings assembled inside procedures are resolved and analyzed instead of skipped, a common blind spot in lineage tooling.
- dbt projects: import the manifest and get column-level lineage across your models.
- Undocumented schemas: the ER diagram view rebuilds table relationships from DDL alone, including idioms like BigQuery’s
NOT ENFORCEDconstraints.
A fair test when you evaluate any online SQL lineage tool: paste your gnarliest stored procedure, not your cleanest SELECT. Open-source parsers such as sqllineage and sqlglot are solid for straightforward single-statement queries; procedural code, dynamic SQL, and dialect edge cases are where a specialized engine earns its keep.
Ask the lineage graph a question
Since version 8.2.3, SQLFlow Cloud includes an AI lineage query. After analysis, ask questions like “which outputs depend on customers.region?” or “summarize how total is computed” in plain English. Unlike a general-purpose chatbot guessing at your schema, answers are grounded: every table and column the AI mentions is validated against the analyzed lineage graph before it reaches you. It is a faster way to interrogate a large graph than clicking through it node by node.
Free vs premium: what $49.99/month buys
| Capability | Free | Premium ($49.99/month) |
|---|---|---|
| Column-level and indirect lineage diagrams | Yes | Yes |
| All 39 SQL dialects | Yes | Yes |
| ER diagrams, call graphs, AI lineage query | Yes | Yes |
| JSON / CSV / PNG export | Yes | Yes |
| Input size | Limited | Larger inputs for real production scripts |
| REST API access | — | Yes — automate lineage in pipelines |
| Team features | — | Yes — share lineage work across a team |
The free tier is a working tool, not a teaser: for ad-hoc questions — “what does this 400-line view actually read?” — it may be all you ever need. Premium matters when SQL scripts outgrow the free input limits, when you want to call the analysis from CI or an internal service through the SQLFlow REST API, or when several people need to work against the same lineage. Full details are on the pricing page.
Run your own SQL through it
The fastest evaluation is your own query in the free tier — no install, no signup for public examples.
When Cloud is the wrong answer
Honest guidance: if your security policy forbids SQL text leaving your network — common in banking, healthcare, and government — use SQLFlow On-Premise instead. It is the same engine packaged for Docker or Kubernetes inside your own infrastructure, runs air-gapped, and costs $500/month or $4,800 one-time per selected database type. On-Premise also covers enterprise scale: batch-scanning estates of 100+ databases and 1M+ columns with incremental scans, a persistent lineage repository, and export adapters for DataHub, Microsoft Purview, and OpenMetadata.
A common pattern: engineers validate the approach on SQLFlow Cloud with representative (non-sensitive) SQL, then the organization deploys On-Premise for the real estate. Nothing learned in the browser is wasted, because the analysis behavior is identical.
Frequently asked questions
Is SQLFlow Cloud free to use?
Yes. The free tier gives you column-level lineage, indirect lineage, ER diagrams, call graphs, and the AI lineage query in your browser. Premium at $49.99/month adds larger inputs, REST API access, and team features.
Does SQLFlow Cloud read my data?
No. SQLFlow performs static analysis of SQL code and schema metadata only; it never reads table row data. If even the SQL text must stay inside your network, choose the On-Premise edition.
Which SQL dialects does SQLFlow Cloud support?
39 dialects with dedicated parsers, including Snowflake, BigQuery, Redshift, Databricks, Oracle, SQL Server, MySQL, PostgreSQL, Teradata, Hive, Spark SQL, Presto, Trino, ClickHouse, DuckDB, and SAP HANA. Each dialect has its own grammar rather than a generic ANSI approximation.
Can SQLFlow Cloud handle stored procedures and dynamic SQL?
Yes. Oracle PL/SQL and SQL Server T-SQL have dedicated procedural parsers; lineage is traced through parameters and temp tables, dynamic SQL inside procedures is resolved, and the procedure-to-procedure call graph is rendered interactively.
What export formats are available?
Lineage exports as JSON (structured, machine-readable), CSV (for spreadsheets and catalogs), and PNG (for documents and reviews). Premium accounts can also retrieve results programmatically through the REST API.
Do I need to install anything to use SQLFlow Cloud?
No. SQLFlow Cloud runs entirely in the browser at sqlflow.gudusoft.com. Installation only enters the picture with the On-Premise edition, which you deploy yourself via Docker or Kubernetes.
Paste one query. See the whole flow.
The fastest evaluation of any lineage tool is running your own SQL through it. SQLFlow Cloud is free to try, right now, in the tab next door.