Checking…
SQL Migration Overview
Total Jobs
0
All time
Analysed
0
Schema ready
SQL Generated
0
Ready to run
Files Processed
0
By Claude AI
Projects
Loading…
No projects yet
Project Status
Live snapshot of the active project
Schema from File
Upload a database export · Claude analyses the schema · download SQL Server CREATE TABLE + INSERT scripts
⚠ Anthropic API key required. Enter your API key from console.anthropic.com — it stays in your browser session only and is never stored on any server.

Job Details
SQL Server Target
🗄️
SQL Server
On-premises
☁️
Azure SQL
Cloud managed
💾
SQL Express
Free edition
🟦
Cloud SQL
Google Cloud
Upload Database Export
Claude reads .sql and text-based files directly. For .bak and .mdb files, metadata and filename are used to guide schema generation.

Drop database files here

.sql · .bak · .mdb · .accdb · .db · .sqlite · .csv

Claude is analysing your database files…
All Migration Jobs
Generated Schema SQL
Generated Migration SQL
Audit Log
Supported File Formats
✓ Full content analysis
.sql MySQL, PostgreSQL, SQLite dumps
.csv Comma-separated data files
.txt Tab or pipe-delimited exports
.json JSON data exports
.sqlite SQLite database files (text mode)
⚡ Schema guidance via filename + metadata
.bak SQL Server backup files
.mdb Microsoft Access databases
.accdb Access 2007+ databases
.db Generic database files
Binary files can't be read directly in the browser. For .bak files, use SSMS to script the schema first and upload the .sql output.
⚙️ Settings
App-level configuration — API credentials, backups, and general preferences. For project metadata (name, client, dates, source/target), open Projects.
💾 Quick backup
Download a full backup of all your jobs, projects, connections and rules.
🔑 API key & credit
Anthropic API key used for all Claude calls. Stored in your browser session only — never sent to any Cygenix server.
No key set
Get a key at console.anthropic.com. Each Claude call (analysis, schema generation, narrative) draws from your account credit there.
Credit status
Live credit balance is shown via the gauge on the New Job page when an API key is set. Anthropic doesn't currently expose an API endpoint for client-side credit polling, so we can't surface the exact balance here without one.
View billing on Anthropic →
🎛️ General preferences
Defaults applied across the app on every page load.
Light mode: soft grey workspace, white panels, dark navy topbar. Financial: FT-style cream workspace with salmon-pink accent and dark masthead. Click Save to persist.
Where Cygenix sends you when you sign in or hit the bare domain.
Used when you click "Export report" without specifying a format.
Auto-snapshots are deduplicated by content hash, so this is cheap.
💾 Backup & Restore
Export all project data to a JSON file, or restore from a previous backup
📦 Export backup
Creates a single JSON file containing:
✓All projects (and active project)
✓Source & target connections
✓All migration jobs
✓Was/Is rules
✓Project artifacts
✓API key
📂 Restore from backup
Upload a Cygenix backup file to restore your project state. Existing data will be merged.
📂
Drop backup.json here or click to browse
🕐 Auto-backup history
Loading…
🔐 Import / Export connections
Move saved source/target connections — including passwords and Azure Function keys — between devices. Connection names sync automatically through your account; secrets stay local by design, and this is how you move them manually.
⬇ Export connections
Pick which saved connections to include, then download a plaintext JSON file.
⚠ Plaintext export. The downloaded file contains database passwords and Azure Function keys in cleartext. Treat it like a credentials file — don't email it, don't commit it, delete it once imported.
⬆ Import connections
Load a previously exported JSON file. You'll see what's in it and can pick which entries to import. Existing entries with the same name will be replaced.
🔐
Click to choose a connections JSON file
🔒 Governance
Control how much of your data reaches Claude during a migration. Settings are stored in your browser (localStorage) and apply across all Cygenix pages that call AI. These controls affect what Cygenix sends to the Anthropic API on your behalf — they do not change anything about Anthropic's own API policies (which state that API inputs are not used to train Claude).
How Cygenix uses AI — plain-English disclosure
⚠ Where Claude currently sees data
  • Dedup AI review — sends row values from MED-confidence clusters to help judge if they're true duplicates (names, addresses, phone numbers, etc.). Only the columns you ticked as match columns, only the clusters that are ambiguous.
  • SQL Editor AI helper — sends whatever SQL you write in the editor, including any literal values in WHERE clauses or INSERT statements.
  • WHERE clause natural-language helper — sends your request plus the table schema (column names + types).
✓ Where Claude does NOT see data
  • AI column mapping (source ↔ target) — only sends schema (column names, types), never row data.
  • One-to-many mapping — only sends schema.
  • Validation — no AI involved, pure SQL execution against your DB.
  • Migration execution — no AI involved.
  • Data insights — schema-only analysis.
Transport: your Anthropic API key lives in your browser's localStorage and is sent directly from your browser to api.anthropic.com. Cygenix servers never see the key or the AI content.
AI access mode
Active: —
🟢 Full
All AI features enabled. Row data and schema can be sent to Claude for mapping, dedup review, and SQL help.
🟡 Schema-only
AI features that would send row data are blocked. Schema-level AI (column mapping) still works. Recommended for client production migrations.
🔴 No AI
All Anthropic API calls disabled. Cygenix works as a pure SQL tool — mapping is manual, no AI assistance anywhere.
Value redaction
Mask values before sending to AI
When AI does see row data (dedup review), mask the values first. Jane Smith → J**e S***h, jane@acme.com → j**e@a**e.com. Claude can still recognise patterns (email, name shape) without seeing the actual content.
Sensitive column exclusion
0 patterns
Columns matching these glob patterns are hidden from AI prompts wherever Cygenix calls Claude — schema introspection, column mapping, dedup review, everything. Use * as a wildcard. Matches are case-insensitive.
AI access audit log
No AI access yet. The log captures every call from a Cygenix page to Claude.
🔌 Integrations
Connect Cygenix to external tools — import/export project bundles, post webhooks on migration events, and more. Credentials are stored in your browser (localStorage) and are not sent to Cygenix servers.
Available now
Working connectors. Configure and use today.
Extensions Available — Contact Sales
Additional connectors available on request. Get in touch to discuss your requirements.
⚙ System parameters
Parameters are tokens like @@StartDate that get substituted with their value wherever they appear — in generated SQL, custom SQL jobs, WHERE clauses, Wasis rules, transforms, and literal values. The code is derived automatically from the name. Pick a Type: number emits unquoted, text and date emit quoted ('ABC', '20260101'), raw passes through verbatim for expressions like GETDATE() or NULL.
Names become codes automatically. Start Date → @@StartDate. Use the code exactly as shown (case-insensitive) wherever you want the value substituted.
🏷
No parameters yet.
Click + Add parameter, or import a CSV/JSON file.
🔌 Project Connections
Configure source and target database connections once — all pages inherit these settings automatically
Supported: SQL Server PostgreSQL
Source Database
Dialect auto-detected from the string. Postgres URLs start with postgres:// or postgresql://
Saved source connections 0
Target Database
Saved target connections 0
🛡 Server Migration
Replicate server-level objects between SQL Server instances. Requires sysadmin on both source and target.
1Verify
2Discover & Select
3Preview SQL
4Execute
5Audit
Verify sysadmin permissions
Server-level operations (creating logins, jobs, deploying SSIS) require sysadmin on both servers. We'll run SELECT IS_SRVROLEMEMBER('sysadmin') against each connection.
Source Click Verify to test connection. Target Click Verify to test connection.
Connections come from the Connections tab. Change them there if needed.

Pick a sheet

This workbook has multiple sheets. Choose which one to import:

Table already exists

The table already exists in the destination. What would you like to do?

➕
Append to existing Insert the new rows alongside current data. Column names and order must match.
⚠️
Overwrite (drop + recreate) DROP the existing table and recreate it from this file. All existing rows will be deleted. Cannot be undone.

Some rows don't fit the chosen types

Review the issues below and decide how to handle them.

🔄
Convert bad values to NULL Keep every row. Values that don't fit the chosen type are stored as NULL. Recommended when a small number of values are dirty.
⏭️
Skip affected rows Drop any row that has at least one bad value. Only clean rows get imported.
🛟
Switch affected columns to TEXT (MAX) Change just the problem columns to NVARCHAR(MAX) and import everything as-is. Escape hatch — fix it up in SQL after.
Connection Diagnostics
Click "Run Tests" to check your Netlify function setup.
Upload & filter
📁
Drop files here or click
CSV · Excel · PDF · Word
SQL · Markdown · Images
Filter by type

Project Artifacts

organised by project folder
✨ Context for Claude AI. Documents uploaded here — emails, client briefs, sign-offs, attachments — are made available to Claude when generating narratives for conversion reports tied to this project. The AI uses them to add context, explain anomalies, and cite sources inline (e.g. [Email, 14 Apr]). Upload artifacts into a project folder and they'll be picked up automatically the next time that project's report runs.
⏱ Task Agent
Schedule migration jobs on cron, run on demand, chain jobs on success
Schedules
Run history
Conversion Reports
Saved runs from Projects → Execute — last 10 per user
After a migration completes, click Save Report to archive it here. Oldest is auto-deleted when you save an 11th.
Report settings
Per-job reconciliation — pick a column to total or count, and Cygenix will compare source vs target on every run.
Showing jobs from your saved project. Changes are saved back to the job record.