Structured data
JSON, JSON Lines, YAML, TOML and XML.
A spreadsheet, a JSON blob, a messy list. PasteCompiler detects what you copied, converts it into the format you need, and shows you what changed.
[
{ "id": "007", "name": "Ada", "role": "design" },
{ "id": "008", "name": "Can", "role": "dev" },
{ "id": "009", "name": "Mina", "role": "ops" }
]Less fixing by hand. More getting things done.
Edit the sample, choose an output format, and copy the result. This small demo runs entirely in your browser.
The browser demo shows 4 output formats. Automatic detection, 17 formats, recipes and detailed conversion reports are in the desktop app.
The desktop app parses supported outputs back and compares them with the input. Rows, columns, cell values and nested keys are checked.
Rows and columnsSee missing rows and changes in structure.
Values and typesSpot values that changed during conversion.
Potential data lossFlattened nesting and other warnings stay visible.
JSON → document → compareprofile.city → "profile.city"round_trip: unavailableSQL, custom templates and plain text do not support round-trip verification. The app explicitly marks it as unavailable in the report.
17 output formats on desktop. Whatever tool comes next, take your data with you.
JSON, JSON Lines, YAML, TOML and XML.
CSV, TSV, Markdown and HTML tables.
SQL INSERT for Postgres, MySQL and SQLite. Optional CREATE TABLE.
Lists, key/value pairs, plain text and custom {{templates}}.
Select columns, filter, sort and deduplicate. Save the steps as a recipe and run them on your next dataset with one click.
↗ Export as a JSON file or a short share code.
Describe a recipe in plain language. Built-in rules work offline; optionally connect Ollama, an OpenAI-compatible server or Anthropic. The built-in rules understand English and some Turkish expressions.
Core conversions and built-in recipe rules run on your computer. If you configure a remote AI provider, you decide whether to send a sample of your data.
Download the ready-to-run Mac app. Core conversions need no Rust, Node.js or AI account.
Get the DMG for Apple Silicon (M1 or newer) and macOS 11+.
Open the DMG and drag PasteCompiler into your Applications folder.
Open the app, paste your data and choose the output format you need.
v0.1.0 is not Developer ID signed or Apple-notarized yet. See the installation guide for the first launch. Installation guide ↗
Yes. PasteCompiler is open source under the MIT license. Core conversions need no subscription. Any remote AI service you choose may charge separately.
The ready-to-run download for this release is only for Apple Silicon Macs. Intel Mac, Windows and Linux users can build from source; this release does not include prebuilt packages for those platforms.
No. Format detection, core conversions, recipes and built-in natural-language rules work without an AI account. Connecting a model is optional.
Core operations run locally. If you connect a remote AI provider, you control data sample sharing in Settings. The small demo on this page also processes data in your browser.
No. Some formats cannot represent the same structure. The app parses supported outputs back to report differences and potential loss. SQL, custom templates and plain text do not support this check.
Yes. Export recipes as JSON files or short share codes starting with pcr1. and import them on another machine.