Build data-driven reports and presentations with AI — without sacrificing accuracy or control.
This package provides Claude with the skills and tools to create professional reports using Motley, where generative AI meets deterministic, auditable data pipelines.
Motley enables number-intensive reporting that combines the ease of use of Generative AI with the robustness and transparency of deterministic configs.
Instead of hoping your AI gets the numbers right, Motley gives you:
- A declarative DSL for charts, tables, queries, and data-driven text — with powerful expressions and cross-references between elements
- A flexible semantic layer for business-relevant queries (use ours, import from your BI tool, or bring your own)
- Tightly controlled LLM calls for text generation that cite their sources and respect your data
- Export anywhere — Google Slides, PowerPoint, and HTML while preserving your original styling
- Full MCP integration so your AI assistant can create and manipulate everything programmatically
The result: reports that are both AI-assisted and trustworthy.
Domain knowledge that helps Claude understand Motley's data model and write correct code:
| Skill | What it does |
|---|---|
create-query |
Build semantic layer queries with measures, dimensions, filters, and time intelligence |
create-edit-chart |
Create bar, line, pie, and funnel charts with customizable styling |
create-edit-text-block |
Generate data-driven text with variable substitution and optional LLM enhancement |
create-edit-table-block |
Build formatted tables with pivoting and flexible layouts |
A complete set of tools for Claude to interact with Motley programmatically:
- Datasource tools — explore and create data sources, add measures and dimensions
- Outline tools — plan deck structure with outline sessions
- Layout tools — browse templates, get thumbnails, create masters
- Master tools — inspect structure, variables, copy/move/delete slides
- Element tools — create and edit charts, tables, text blocks, queries
npx skills add MotleyAI/agent-skills -a claude-codeDownload the .mcpb file from Releases and drag it into Claude Desktop.
export MOTLEY_API_KEY="sk_your_key_here"
claude --plugin-dir /path/to/agent-skillsSee docs/setup.md for detailed instructions.
flowchart LR
subgraph AI["🤖 AI Assistant"]
Claude["Claude"]
Skills["Skills"]
MCP["MCP Tools"]
end
subgraph Motley["⚙️ Motley Engine"]
DSL["Declarative DSL"]
Semantic["Semantic Layer"]
end
subgraph Export["📄 Output"]
Slides["Google Slides"]
PPT["PowerPoint"]
HTML["HTML"]
end
subgraph Data["💾 Your Data"]
DB[(Database)]
BI["BI Tools"]
end
Claude --> Skills
Claude --> MCP
Skills --> DSL
MCP --> DSL
DSL --> Semantic
Semantic --> DB
Semantic --> BI
DSL --> Slides
DSL --> PPT
DSL --> HTML
style AI fill:#e8f4f8,stroke:#0891b2
style Motley fill:#fef3c7,stroke:#d97706
style Export fill:#dcfce7,stroke:#16a34a
style Data fill:#f3e8ff,stroke:#9333ea
sequenceDiagram
participant You
participant Claude
participant Motley
participant Data
participant Export
You->>Claude: "Create a quarterly revenue report<br/>with regional breakdown"
Claude->>Claude: Uses skills to write<br/>correct Motley DSL
Claude->>Motley: Creates report via MCP tools
Motley->>Data: Executes queries<br/>deterministically
Data-->>Motley: Returns results
Motley-->>Claude: Preview ready
Claude-->>You: "Here's your report"
You->>Motley: Export
Motley->>Export: Google Slides /<br/>PowerPoint / HTML
Every number is traceable. Every chart is reproducible. Every report is auditable.
- Setup Guide — Installation and configuration
- Skills Reference — Domain knowledge for queries, charts, text, and tables
- MCP Tools Reference — All 32 MCP tools with detailed argument documentation
- Website: motley.ai
- Questions? Open an issue or book a demo
MIT