High-performance local data workspace
High-performance viewing and editing for large CSV files, quantitative research and AI training data. A native Rust engine does the heavy local work—so you can finish the task and get back to your day.
Free Preview · up to 100 records · signed in-app updates · Install guide →
| # | symbol | timestamp | close | volume |
|---|---|---|---|---|
| 1 | DEMO-A | 00:00:00 | 101.25 | 320 |
| 2 | DEMO-B | 00:00:00 | 202.50 | 640 |
| 3 | DEMO-C | 00:00:00 | 303.75 | 480 |
| 4 | DEMO-A | 00:01:00 | 101.50 | 350 |
| 5 | DEMO-B | 00:01:00 | 202.75 | 610 |
| 6 | DEMO-C | 00:01:00 | 304.00 | 500 |
For people working at data scale
Training data, market history or a multi-gigabyte export: see the records that answer the question, with their source context intact.
AI training & data engineering
Pretraining shards, nested conversations and preference pairs. Inspect JSONL, Parquet and Arrow without turning the entire corpus into a desktop table.
AI training data workflowsQuantitative research
Large market-data exports, trade histories and parameter sweeps. Narrow symbols and time ranges, read exact values and trace findings to source records.
Market data & backtest workflowsLarge-file analysis
GB-scale CSV, JSONL and columnar files. Start with a useful page, query supported sources and save reviewed edits as a separate copy.
Large-file viewing & editingPerformance that serves the workflow
Native Rust handles decoding, indexing, querying and persistence. Supported column projection and partition pruning reduce the data read; virtualized tables keep the desktop view focused on visible records.
Open a bounded first page, narrow the query, inspect the exact value and move on. Large scans keep progress and cancellation visible.
How the performance model worksPurpose-built data workflows
Choose a synthetic example to explore the workflow. The illustration shows field selection—not measured speed or guaranteed byte savings.
Inspect the complete value, narrow the source query, save a reviewed copy.
Synthetic field counts. Not a performance benchmark.
| record_id | category | amount | status |
|---|---|---|---|
| mock_001 | Demo A | 125.00 | review |
| mock_002 | Demo B | 98.50 | ready |
| mock_003 | Demo A | 210.25 | review |
CSV synthetic example selected
One workspace, many formats
Text exports, training records, columnar shards and local databases—with the reader and available operations made explicit.
Local data, under your control
Parsing, indexing, queries and supported edits run locally in the Rust data engine. Dataset upload is not required to inspect your files.
Account activation and updates use network services. MCP sends the requested results to the AI client you choose; a cloud-backed client may pass them to its model provider.
Understand the privacy boundary →MCP sharing is a separate, explicit connection.
The point of faster data work
Fewer temporary scripts. Less manual file splitting. Less staring at a frozen table.
Spend your attention on the finding—and leave room for dinner with family, a walk outside or a game with friends.
Find the workflow that saves you steps →Inspect the training-data schema.
Choose one split and partition.
Return only the required fields.
Include source-file provenance.
Do not edit or export yet.AI-assisted data work
Use authorized local MCP tools for discovery, schema inspection and supported data queries. Review schemas, test calls and wire logs in the desktop client.
Set up a bounded MCP workflowA tool that earns back your time
Test a bounded first page on your own file. Pro adds supported full-file queries, quality checks, editing, export and local MCP workflows.
Preview
No account. No card. Start with your own file.
Get PreviewPro
For large-data workflowsOr $150 / year — save $30 annually.
Renews automatically. Cancel renewal anytime. Billed upfront for the selected period. Final tax and total at checkout.
RowCove is a local data workspace for inspecting, querying and safely editing data files on your computer. Its native Rust data engine handles local parsing, indexing, querying and persistence for large-file workflows.
The desktop data engine processes files locally. Account services receive your email and purchase status, not your datasets. If you connect an AI client through MCP, the tool results you request are shared with that client and may reach its model provider.
Preview is free and does not need an account. Inspect supported file information, schema and up to 100 records in a read-only first page. Pro adds full-file workflows and MCP; available operations depend on the reader.
Pro costs US$15 per month or US$150 per year. The yearly plan is billed upfront and saves US$30 compared with twelve monthly payments. Applicable taxes and the final total are shown at checkout.
Yes. Open Account and choose Manage subscription. Cancel renewal in the Paddle portal. Access continues through your paid period; approved full refunds revoke the refunded purchase.
Activate Pro online and reconnect periodically to keep access current. An active subscription is required.
Windows is the currently validated desktop platform. macOS release acceptance is still pending. Check the release page for published builds.
No. Reader capabilities differ. CSV, JSONL, Parquet and Arrow have table workflows; database sources have explicit table selection. Some scientific and training formats currently expose metadata or archive members rather than decoded samples. See the format guide.
AI data engineers working with pretraining and post-training corpora, quantitative researchers with large market or backtest data, and people inspecting or editing GB-scale CSV, JSONL and columnar files.
TB-scale datasets are a target use case for scoped access through supported shards, partitions and projected fields. A TB-scale throughput and memory benchmark has not been established. Test a representative subset and plan scan, decompression and scratch-disk costs first.
Rust handles parsing, indexing, querying and persistence locally. Supported projection and partition pruning reduce unnecessary reads, while bounded results and virtualized views keep the desktop interface focused on the records you need.
Yes. Pro provides authorized local MCP tools for supported discovery, schema and data workflows. Review tool schemas, test calls and logs in the desktop client. Requested results go to the connected AI client and may reach its model provider.
No. All product screenshots, examples and benchmark inputs use generated synthetic (mock) data. Comparison timings are measurements on mock inputs.
Built for the file you have been putting off
Start with a bounded first page on your own machine.
Windows Preview available. macOS release acceptance pending.