How waqasdaar/chartiva works
Chartiva is a commercial MCP server and agent skill that lets AI assistants turn user data and plain-language requests into interactive visualizations and analyses. The documented flow is: the assistant uses server tools to profile data, choose and render a chart, check it, and return browser-ready HTML or other exports. Separate capabilities include dashboards, linked datasets, network analysis, anomaly detection, and conversational revision. The repository tree contains only showcase assets and a demo, so implementation paths and internal call relationships cannot be verified; edges below reflect README statements only.
Components 18
Assistant interface
- AI assistant
- Agent skill
- MCP server
Data analysis
- Data ingestion
- Data profiling
- Chart planner
- Dataset linking
Visualization
- Chart renderer
- Result checking
- Interactive HTML
- Chart exports
- Dashboards
Analysis features
- Network analysis
- Anomaly detection
- Chart revision
- Version history
Other components
- User
- User data
Connections 18
- User → AI assistant: asks questions
README.md - User → User data: provides data
README.md - AI assistant → Agent skill: follows guidance
README.md - AI assistant → MCP server: uses tools
README.md - MCP server → Data ingestion: reads formats
README.md - User data → Data ingestion: supplies input
README.md - Data ingestion → Data profiling: profiles data
README.md - Data profiling → Chart planner: informs choice
README.md - Chart planner → Chart renderer: selects chart
README.md - Chart renderer → Result checking: checks result
README.md - Result checking → Interactive HTML: returns HTML
README.md - MCP server → Chart exports: exports formats
README.md - MCP server → Dashboards: creates dashboards
README.md - MCP server → Dataset linking: links datasets
README.md - MCP server → Network analysis: computes analytics
README.md - MCP server → Anomaly detection: detects anomalies
README.md - AI assistant → Chart revision: refines conversationally
README.md - Chart revision → Version history: tracks versions
README.md