The semantic layer & reporting engine built for AI agents.
Motley gives AI agents a governed view of your data and turns their output into repeatable, forwardable reports. Build on the open-source semantic layer (SLayer), generate reports from any MCP client, or embed the whole stack in your product.
Trusted by:
What is Motley
SLayer (open source)
The semantic layer that makes AI agents actually understand your data.
SLayer sits between your warehouse and your agents. Agents query in terms of measures, dimensions, filters, and time grains. SLayer compiles to SQL, runs it against the warehouse, and returns typed results. One definition of every metric, enforced across every agent and app, less silent failures
Key points:
-
Agent-first design: MCP, REST API, and Python SDK
-
Rich metric expressions: reusable measures, dimensions, filters, and logic
-
Embeddable semantic layer: use via API, plugin, or inside your product
-
Datasource-agnostic: Postgres, Snowflake, BigQuery, ClickHouse, and more
-
MIT licensed and extensible
View on GitHub
Motley Plugin
Connect your AI agent to governed data
Motley exposes your data to AI agents via MCP. The plugin uses SLayer as a semantic layer, ensuring consistent metrics, joins, and definitions across queries. Ask for a report in natural language – Motley generates the queries, structures the document, and returns production-ready output.
Key points:
-
MCP interface for Claude and compatible AI agents
-
Query governed metrics defined in SLayer
-
Connect to SQL databases, BI tools, and APIs
-
Outputs structured documents: PPTX, docs, dashboards
-
Iterative workflow via conversational interface
Try the Plugin
Embedded Motley
Reports inside your product, powered by Motley.
Ship AI-powered reporting in your product without building the stack yourself. Motley runs server-side behind your UI: your customers get branded reports that regenerate as their data changes, you get a single integration instead of a six-month build.
Key points:
-
White-label: your brand, your interface, Motley under the hood
-
Live data: reports update when the underlying data changes
-
Multi-tenant: each customer sees only their data
-
Flexible integration: iframe, JS snippet, or headless API
-
Flexible integration: iframe, JS snippet, or headless API
Learn About Embedded
Integrations
How It Works
Connect Your Data
Plug in your database, data warehouse, or BI tool. SLayer introspects the schema and generates semantic models automatically — no manual configuration needed
Define What Matters
Refine your models with business-specific metrics, dimensions, and context. This is where organizational knowledge becomes machine-readable. Define it once; every agent and app uses the same definitions.
Generate Reports
Use the Motley Plugin to create reports conversationally, or call the API programmatically. Motley pulls data through SLayer, builds the narrative, and outputs a finished deck, document, or dashboard.
Ship It
Deliver reports as PPTX, embed them in your product with Embedded Motley, or build custom experiences on top of the SLayer API. Your data, your story, your way.
The Problem
Before Motley
Agents hallucinate metrics because they're writing SQL against a schema they don't understand
Every team defines "ARR" or "active user" slightly differently, and no one notices until the board deck
Reports are one-offs – rebuilt by hand every cycle, impossible to audit
Shipping reporting in your product means staffing a data platform team
Dashboards answer "what happened" but not "why it matters" — and no one reads them
With Motley
Agents query through a governed semantic layer, not the raw schema
One definition of every metric, enforced across every agent and surface
Reports are markdown files: versioned, diffable, forwardable, regenerable
Embedded Motley ships reporting into your product without a platform build
Narrative-first output: rose, charts, and queries in one artifact
Whoʼs Motley for
Data & Engineering Teams
Use SLayer as the governed data access layer for AI agents. Define metrics once, query from anywhere — MCP, REST, Python, or CLI.
Product Teams Building Reporting
Add AI-powered reporting to your SaaS product without building a data pipeline. Embedded Motley handles the data layer, narrative, and presentation.
Revenue Operations
Generate pipeline reviews, forecast decks, and performance summaries from live CRM and database data.
Customer Success Teams
Automate QBRs, monthly check-ins, onboarding updates, and value reports. Spend your time on retention and growth, not slide assembly
"Our challenge is that none of the existing AI tools seems to do the data crunch, visualisation and consistent report building at the same time. ChatGPT is awfully inconsistent, Gamma can't crunch the data and Julius doesn't have a repeatable process"
Erik Beentjes
FAQs
Motley is an open-source semantic layer (SLayer) and a reporting engine built on top of it. Agents query your data through SLayer; Motley turns those queries into reports you can forward, version, and embed.
SLayer is agent-native: it exposes 14 tools over MCP out of the box, supports composable time formulas, and ships with pluggable storage. It's designed to sit in front of AI agents, not BI dashboards.
Text-to-SQL hallucinates joins, silently redefines metrics, and produces different numbers on different runs. A semantic layer constrains the agent to a governed set of measures and dimensions: same query, same answer, every time.
SOC 2 Type 2 on track for 2026. Data shared with model providers is scoped to the report being generated, never used for training, and can be self-hosted end-to-end if you run SLayer yourself.
Data and engineering teams that want to expose their warehouse to AI agents safely. Product teams shipping reporting features. Developers who want their agents to generate accurate reports instead of guessed SQL.
SLayer is MIT-licensed and free. The Motley plugin has a free tier (10 docs/month) and paid plans from ~$5/month. Embedded Motley is priced per design partner — contact us.
Reports, not dashboards. Markdown as the source of truth. A semantic layer you can actually read the code of. Built for the agentic era from the first commit.