Argument Map Editor
Map what supports a position, what challenges it and which assumptions it depends on. Add a source title, optional HTTP(S) link and page or table locator to evidence and objections. The directed graph, full-text outline and warnings remain linked while you edit. Warnings identify structural review points; they cannot decide whether an argument is true.
Key features
- Four distinct node roles: claim, evidence, counterargument and assumption
- Directed support, attack and dependency links with editable relation types
- Warnings for claims without incoming support, isolated nodes and missing source titles
- Reasoning-cycle detection across support and dependency links
- Visual graph plus complete keyboard-operable outline and relation list
- Strict editable JSON import and JSON, CSV report and standalone SVG exports
How to use
- Load the fictional example or edit the project title and selected statement.
- Add typed claims, evidence, counterarguments or assumptions and cite sources where available.
- Select a node and draw a directed support, attack or dependency relation to another node.
- Inspect the graph, full-text outline and warnings; revise missing sources or cycles.
- Download JSON to continue later, CSV for review or SVG for a static diagram.
Use cases
- Review a policy proposal with its supporting evidence and objections
- Identify an assumption shared by several claims in a team discussion
- Prepare a research briefing that shows uncited evidence and unsupported claims
Frequently asked questions
What do the arrows mean?
A relation points from the selected source node to its target. Supports and attacks describe the selected statement's effect on the target; depends means the selected statement relies on the target. The arrow labels remain editable in the relation list.
When is a claim called unsupported?
A claim or counterargument receives this warning when no direct incoming Supports arrow exists. It is a completeness check, not a judgment that cited evidence is sound or the statement is false.
Which cycles are reported?
A directed cycle using Supports and Depends links is flagged as circular reasoning to review. Attack-only cycles are debate links and are not treated as proof cycles. Cycles are warnings, so you can inspect and repair them without losing the project.
Which nodes need a source?
Evidence and counterarguments without a source title are flagged. A source title can identify a book or offline document; an optional HTTP(S) URL and locator provide more precise citation. The tool does not fetch or verify the source.
What happens when I delete or import?
Deleting a node removes only its incident relations. A valid version 1 JSON import replaces the current project; malformed UTF-8, duplicate keys or IDs, broken endpoints, extra fields and unsafe URLs are rejected without changing it.
Does the page save or share my argument automatically?
No. Editing stays in this tab's memory. Refresh resets to the example, so download the editable JSON before leaving. CSV and SVG are for review and presentation, not reimport.
Privacy
Statements, citations and imported files are processed in browser page memory. This tool does not fetch source URLs, autosave or upload the project. Downloads occur only when you choose them; inspect exports before sharing.
Comments & questions