Settings

Tune AutoPilot to match your hardware

Performance

Adjust processing limits based on your computer's RAM. Higher values unlock larger datasets and faster processing but require more memory.

Quick Presets

Custom Limits
Fine-tune individual settings. Changes take effect on the next workflow run.
10,000
5K — less RAM, more files100K — more RAM, fewer files

What this does: Controls how many rows are written per chunk file in OPFS storage. Larger chunks mean fewer file operations (faster on SSD) but each operation uses more RAM. Increase this if you have 8GB+ RAM and want faster CSV processing.

500,000
100K — safe on 4GB RAM10M — needs 16GB+ RAM

What this does: Sets the in-memory row threshold used by the CSV Join and Compare nodes. Below this limit both nodes use a fast in-memory Map; above it they switch to a disk-based grace hash join via OPFS to avoid running out of RAM. Roughly 1 GB of free RAM per 2 million rows.

How to choose the right settings for your PC
4 GB RAM (low-end laptop)→ Use Balanced preset→ Chunk: 10K, Union limit: 500K→ Max recommended CSV: ~200MB
8 GB RAM (modern laptop)→ Use Performance preset→ Chunk: 25K, Union limit: 2M→ Max recommended CSV: ~500MB
16 GB RAM (workstation)→ Use Maximum preset→ Chunk: 50K, Union limit: 5M→ Max recommended CSV: ~1GB
32 GB+ RAM (high-end)→ Use Custom sliders→ Chunk: 100K, Union limit: 10M→ Max recommended CSV: 2GB+

Note: These limits apply per workflow run. If a workflow fails with an out-of-memory error, lower the chunk size and union limit, then try again. Settings are saved locally and persist between sessions.

Storage Management
Execution data (CSV chunks) is stored in your browser's private file system (OPFS). It is never uploaded anywhere but accumulates over time.
Browser storage used—
Saved executions—run records in IndexedDB
Saved workflows—workflow definitions
Stored datasets—CSV chunk manifests
AutoPilot v0.1.0 — offline-first, all data stays in your browser