What a Key Adds
This page lists exactly what lexega-sql does without a key and what needs a trial or Team key. It also shows a finding that only the key produces, and how to tell from any report which of the two ran.
The split#
| Feature | Without a key | With a trial or Team key |
|---|---|---|
Formatting (fmt) | Yes | Yes |
analyze, review and ci | Each statement as written | Also across statements, scopes, procedures, dbt models and scripts |
Comparing the two versions of a changed statement in review | No | Yes |
The diff command | Refused | Yes |
--cross-script | Refused | Yes |
dbt and Jinja rendering on analyze, review, ci, fmt and mcp | Yes | Yes |
Policy enforcement and decision records: --policy, --exceptions, --decision-out | Yes | Yes |
The dashboard and mcp commands, analyze --mode runtime | Yes | Yes |
| Every built-in rule and dialect, custom rules | Yes | Yes |
--catalog and catalog pull | Yes | Yes |
| Cloud storage locations for rules, reports and artifacts | Yes | Yes |
JSON, YAML, Markdown, SARIF and GitLab SAST output, --pr-comment | Yes | Yes |
--strict, init, policy init, policy-lint | Yes | Yes |
"Refused" means the run stops before it analyzes anything, names the feature that needs a key, and exits 1.
Depth is the one difference that is not refused. Without a key, analyze, review, ci and mcp run every rule, and the rules that need analysis beyond a single statement as written report what they can and stay silent where they would need more. A dbt model is rendered and then analyzed on its own; with a key it is also analyzed with the models it reads.
A policy decides on the findings the run has, so without a key it decides at that depth. The decision record carries the same analysis_depth a report does, and an answer from the mcp server carries an analysis_note.
The depth note#
Every report from a run without a key says how many rules that affects. In the text report it is the Note: line under Confidence::
Note: 57 of 957 rules use analysis that needs a license key and may stay silent or report less precisely.
The second number is the number of rules the run evaluated, so it grows with your custom rules. With a key, the line is not printed.
The same file, without and with a key#
A LEFT JOIN keeps every customer, with or without orders. The WHERE condition on o.status then throws away the rows with no order, because their o.status is null, so the query returns only customers who have a shipped order. Save this as orders_report.sql:
SELECT c.customer_id, c.name, o.order_id, o.total
FROM customers c
LEFT JOIN orders o ON o.customer_id = c.customer_id
WHERE o.status = 'shipped';
lexega-sql analyze orders_report.sql
Without a key:
Analysis Report
========================
Confidence: HIGH (full render, no placeholders)
Note: 57 of 957 rules use analysis that needs a license key and may stay silent or report less precisely.
Summary:
Total findings: 0
Critical: 0
High: 0
Medium: 0
Low: 0
Context: 0
Coverage:
Analyzed: 1 SQL statements
Data Operations:
Tables Read: 2
Cross-Database: No
With a key:
Semantic Analysis Report
========================
Confidence: HIGH (full render, no placeholders)
Summary:
Total findings: 1
Critical: 0
High: 1
Medium: 0
Low: 0
Context: 0
Coverage:
Analyzed: 1 SQL statements
Data Operations:
Tables Read: 2
Cross-Database: No
findings:
[HIGH] Outer join's nullable side filtered in WHERE clause or by a later join. The unmatched rows the outer join keeps are dropped, likely a bug.
↳ Line 4 • `Q-JOIN-LEFT-FILT` • "status" • orders_report.sql:4:9
Both runs read the same statement. Finding Q-JOIN-LEFT-FILT needs to know which side of the join can be null and that the WHERE condition rejects a null, which is analysis across the query's parts. The run without a key reports nothing for this file, and its Note: line is the only sign that a rule like this one could not fully run.
Tell which depth you ran#
Each output format carries the depth note when a run was made without a key, and leaves it out with a key:
| Format | Where to look |
|---|---|
| Text | The Note: line under Confidence: |
Markdown, including review and --pr-comment | A **Note**: line under the summary |
| JSON and YAML, one file | summary.analysis_depth, with rules_total and rules_limited |
| JSON and YAML, a folder or a review | batch_summary.analysis_depth, and the summary of each file's report |
| SARIF | A notification with the id analysis-depth under runs[].invocations[].toolExecutionNotifications |
| GitLab SAST | A message under scan.messages |
In JSON, the run without a key above has this in its summary:
"analysis_confidence": "high",
"analysis_depth": {
"rules_total": 957,
"rules_limited": 57
}
The run with a key ends its summary at "analysis_confidence": "high". A script that must not accept a reduced-depth report can fail when analysis_depth is present.
Trial, Team and Enterprise#
A trial key unlocks every feature for 30 days: start a trial. A Team key unlocks the same features. Enterprise has the same features as Team, with support terms by agreement. Pricing has the details, and Quick Start shows how to add a key on a machine or in CI.
Something missing or wrong? Open an issue on GitHub or write to [email protected].