A structured document editor for composing meaning in JSON.
jsonedit is a tree-based editor designed for working with hierarchical ideas, not just JSON text. It treats JSON as a living document — something you assemble, reshape, and extract from — rather than a file you manually type.
It is especially well suited for LLM prompt construction, context curation, and structured knowledge editing.
jsonedit is not a traditional programmer’s JSON editor.
It is a document compositor built around one core idea:
Edit meaning locally. Commit structure deliberately.
Instead of editing lines of text, you work directly with the document’s tree structure — selecting nodes, reshaping hierarchies, and exporting semantic fragments when needed.
Modern workflows — especially those involving LLMs — often look like this:
- assemble structured prompts
- rearrange examples or schemas
- experiment safely
- extract only the relevant subtree
- paste into an AI context window
- iterate
Traditional editors treat JSON as text.
jsonedit treats JSON as thought structure.
The primary editing unit is a node, not a line.
Common operations include:
- raise / lower items in hierarchy
- duplicate subtrees
- insert siblings
- rename keys
- delete with intelligent reselection
Navigation uses structural paths instead of line numbers.
Two coordinated views:
- Tree view → orientation and structure
- Text pane → focused semantic rewriting
You always know where you are structurally while editing what something means.
Text edits remain sandboxed until you explicitly commit them.
This allows:
- experimentation without breaking structure
- temporary invalid JSON while thinking
- deliberate structural mutation
No accidental global edits.
jsonedit assumes you frequently want pieces of documents.
You can quickly copy:
- entire documents
- selected subtrees
- pretty or compact JSON
Ideal for moving structured context into LLM prompts.
Documents may include metadata and workspace information, encouraging long-lived structured artifacts rather than disposable config files.
jsonedit works best when JSON represents ideas, not just data.
- LLM prompt composition
- Context packaging for AI systems
- Agent or tool configuration
- Structured writing and outlining
- Dataset annotation
- Knowledge assembly in hierarchical form
Typical workflow:
- Select subtree
- Rewrite locally
- Commit structure
- Copy compressed node
- Paste into LLM context
- General programming JSON editing
- Schema validation or linting
- Large machine-generated JSON files
- Collaborative merge workflows
- High-speed code typing environments
If you want VS Code for JSON — use VS Code.
jsonedit optimizes for clarity of structure, not typing speed.
pip install jsoneditjsoneditOpen a file directly:
jsonedit <filepath>- Python
- Tkinter available in your Python installation
(jsonedit is a desktop GUI application.)
jsonedit is built on a few assumptions:
- JSON encodes ideas, not just data.
- Structure matters more than formatting.
- Users think in subtrees rather than lines.
- Editing should be deliberate and reversible.
- Exporting fragments is a primary activity.
CC0 1.0 Universal — public domain.
See LICENSE.