Eleven databases.
One SQL tool.
The old paradigm is out. The new paradigm is AI. AI automation is InTouch AI — and one AI-native engine runs every database you have. A specialized DB scheduler can do one thing; a general engine does all of it, and the reverse is impossible. Drivers for Oracle, SQL Server, MySQL, PostgreSQL, DB2, Informix, MariaDB, Derby, Cassandra, Firebird, and Cloud Spanner all ship in every edition — including the free Personal edition. Write a query once, target any of them via a credential reference. Batch inserts, large-result streaming, JDBC pool management, and a proprietary batch pathway that delivers orders-of-magnitude faster imports for bulk loads.
All Common Enterprise Stacks, Plus Cassandra and Spanner
One tool, every engine. No per-database point product to license, learn, and babysit. A config-era tool is welded to the database it was built for. InTouch AI is not — the same SQL task reaches all of them through a named credential.
Oracle
11g through 23c. Connection tested against the full supported matrix.
SQL Server
2012 through 2022, plus Azure SQL Database.
MySQL
5.7 and 8.x, MySQL-on-RDS, and drop-in-compatible forks.
PostgreSQL
12 through 17, including Aurora PostgreSQL.
IBM DB2
LUW and z/OS. The databases that run the Fortune 500's back office.
Informix
Legacy-but-still-here databases, fully supported — no archaeology required.
MariaDB
10.x series. Drop-in compatible with MySQL queries.
Derby
Embedded default. The database the Personal edition ships with — zero config, production-grade.
Cassandra
CQL via the DataStax driver. Wide-column at scale.
Firebird
2.5 and 3.x. Small-footprint RDBMS for edge deployments.
Cloud Spanner
Google's global-consistency SQL database, via the official JDBC driver.
MongoDB
NoSQL via the MongoDB plugin — separate from the SQL tool.
Common Database Automation Patterns
Every pattern carries the same contract, now intelligent: tell it what to do, when, what to do when it works, what to do when it fails, and who to notify. The "when it fails" clause is no longer a dumb retry-and-email-a-log. The engine reads the failure, knows why, refreshes the expired DB credential, smart-retries the deadlocked load, and surfaces the one line that matters. It broke. Here's why. I fixed it. No config-era scheduler can say that.
ETL / ELT
Extract from one database, transform with the DataFrame tool, load into another. Schedule on the seven-type scheduler, notify on completion across whichever messaging channels you've configured (Email, Slack, Discord, Telegram, and SMS on every edition; WhatsApp, Microsoft Teams, and LINE add on Department/Enterprise), audit every row count.
Bulk Import
The proprietary batch pathway delivers up to 1000x faster imports than row-at-a-time inserts. Large CSV to Oracle in minutes, not hours.
Scheduled Reports
Query → DataFrame → format as XLSX or PDF → email to recipients. A single job, seven schedule types to pick from, zero glue code.
Data Quality Monitoring
Compare row counts and hashes across environments. Alert when drift exceeds thresholds. A Monitor with an ai: arm can escalate the weird ones to humans.
Every Database Job Is a First-Class Object
Database credentials live in the AES-256 credential vault — referenced by name, never written into a script, never exposed even to the AI itself. That is the trust floor under every query. Queries are stored as first-class job objects with owner, group, and per-role rights. Every execution writes an audit record with start time, end time, row counts, and outcome. Failures page whichever team you configured, through whichever messaging channel they prefer (Email, Slack, Discord, Telegram, and SMS on every edition; WhatsApp, Microsoft Teams, and LINE add on Department/Enterprise). This governance was hardened in Fortune 500 production for 25+ years — before "AI automation" was a phrase.
Your Databases, Automated
One AI-native engine for all eleven. Download the free Personal edition. Connect to your databases. Write your first job in about 15 minutes — and let it heal its own failures from there.