The biggest release since launch. Datavor now ships with a built-in scheduler, a transform pipeline, a real-time sync dashboard, and an intelligent type engine β all free, all conversational.
When we shipped Datavor v1.0, we proved one thing: you could connect two databases and sync a table just by talking to Claude. No SQL. No YAML config files. No paid SaaS subscription.
What we couldn't do yet was keep those syncs running on their own, transform data as it moved, or give you visibility into what was actually happening inside your pipelines. Those were the gaps. v1.5 closes all of them.
npm install -g datavor@latest and restart Claude Desktop. Your existing connections and sync history carry over automatically.This release completes what we're calling Phase 2 of Datavor's development. Four new capabilities, built and tested against live MySQL and PostgreSQL databases β local and cloud (Aiven).
Create, pause, resume, and delete recurring sync jobs with natural language. Cron-style scheduling under the hood β managed entirely through Claude. No external cron daemon required.
Apply column mappings, data type conversions, renames, and field-level transformations inline during a sync β before data lands in the destination. Preview changes on real rows before committing.
Real-time visibility into every pipeline: sync history, row counts, durations, failure diagnostics, and overall health. Ask Claude for a summary and it reads straight from the dashboard.
Intelligent type mapping between MySQL and PostgreSQL schemas. Handles edge cases like TINYINT(1) β BOOLEAN, DATETIME β TIMESTAMP, and TEXT variants. SSL support for cloud-hosted instances.
The most-requested feature since launch. You can now tell Claude to run a sync on a schedule, and it just works. Jobs persist across sessions and Claude Desktop restarts. You can list running jobs, pause one while keeping others active, or fire a job manually outside its schedule.
Data rarely moves cleanly between systems. Column names differ. Types don't match. Enum codes from the legacy system mean nothing in the new one. The Transform Pipeline lets you describe these changes in plain English and applies them inline during sync β no intermediate staging table needed.
You can also call transform_preview to see exactly what the transforms will do to a sample of your real data before the sync runs. No surprises.
The dashboard gives you a live view of everything Datavor has done and is doing. Ask for a summary and you get total syncs, success rates, row volumes, and the most recently active tables. Drill into a specific table's history to see every run β status, rows moved, duration, and error messages when things go wrong.
MySQL and PostgreSQL have subtly different type systems. Moving data between them used to mean guessing at mappings or watching syncs fail silently when types didn't line up. The Type Engine handles this automatically, mapping every MySQL type to its correct PostgreSQL equivalent, with warnings surfaced to Claude when an ambiguous case is detected.
v1.5 also ships with SSL support for cloud-hosted databases, validated against Aiven-managed MySQL and PostgreSQL instances. If you're syncing between cloud databases, it just works.
Let's be direct about what this means competitively. Enterprise data integration platforms typically charge between $239 and $2,000+ per month for the capabilities Datavor now delivers for free. The table below shows where those costs come from.
These are excellent platforms built over many years with extensive feature sets. Datavor isn't trying to clone them β it's doing something different: bringing core ETL capabilities into the AI-native, conversational layer. For solo developers, startups, and small teams managing pipelines on a Mac, that's a fundamental shift.
Datavor 3.0 β with its expanded connector library and advanced pipeline management β is designed to compete feature-for-feature with tools like Hevo Data and integrate.io in their core data sync use cases. We're not there yet. But v1.5 is the foundation, and it's solid.
MySQL β PostgreSQL sync, full and incremental modes, connection management, schema inspection. β Shipped.
Scheduler, Transform Pipeline, Sync Dashboard, Type Engine, SSL support, full test suite. β Shipping today.
SQL Server, SQLite, Snowflake. Windows and Linux support. Bidirectional sync improvements.
Browser-based dashboard, visual pipeline builder, team features, MongoDB support. Competing feature-for-feature with Hevo Data and integrate.io.
Datavor runs locally on your Mac as an MCP server connected to Claude Desktop. There's no account to create, no API key to manage, no SaaS dashboard to learn.
Once connected, Claude gains access to all of Datavor's tools. Connect your first database, browse your schema, run a sync β all in natural language. The full documentation is at datavorai.com/docs.html.
Datavor started as a frustration: why does moving data between two databases require a SaaS subscription, a DevOps pipeline, and a senior engineer to set it up? It shouldn't. Data engineering shouldn't be gated behind enterprise budgets.
The MCP protocol β Anthropic's open standard for connecting Claude to external tools β gave us the right foundation. By building Datavor as an MCP server, we get to slot into Claude Desktop with a single config change. No new interface. No browser tab. Just Claude, doing more.
v1.5 is the most complete version of that vision yet. We hope you find it useful.