Context
Suggested by Gemini 2.0 during LLM evaluation. A visual interface dramatically lowers the barrier to adoption.
Problem
Currently, KCP is entirely CLI/API-driven. There's no way for non-technical users (or technical users who prefer a GUI) to:
- Browse their knowledge base
- See the lineage DAG visually
- Search with a text input
- Inspect artifact signatures and metadata
Proposed Solution
A single-file ui/index.html served by kcp serve at http://localhost:8800/ui.
Features
Knowledge Browser
- List all artifacts with title, tags, date, format badge
- Full-text search with highlighting
- Filter by: tag, format, visibility, date range
Lineage Visualizer
- Interactive DAG using D3.js or Cytoscape.js
- Click a node to see artifact details
- Highlight derivation chains
Artifact Inspector
- View full content (markdown rendered)
- Show cryptographic signature + content hash
- Copy
artifact_id to clipboard
- One-click
kcp_get / kcp_lineage output
Stats Panel
- Total artifacts, by format, by tag
- Storage size
- Last 5 published
Implementation Notes
- Single HTML file (no build step, no npm) — same philosophy as KCP itself
- Uses
fetch() against the local kcp serve HTTP API
- Works offline (all local)
Acceptance Criteria
Related
sdk/python/kcp/cli.py (serve command)
- Suggested by: Gemini 2.0
Context
Suggested by Gemini 2.0 during LLM evaluation. A visual interface dramatically lowers the barrier to adoption.
Problem
Currently, KCP is entirely CLI/API-driven. There's no way for non-technical users (or technical users who prefer a GUI) to:
Proposed Solution
A single-file
ui/index.htmlserved bykcp serveathttp://localhost:8800/ui.Features
Knowledge Browser
Lineage Visualizer
Artifact Inspector
artifact_idto clipboardkcp_get/kcp_lineageoutputStats Panel
Implementation Notes
fetch()against the localkcp serveHTTP APIAcceptance Criteria
sdk/python/kcp/ui/index.html— single-file web app/uibykcp serveRelated
sdk/python/kcp/cli.py(serve command)