llms.txt
# Marmot
> The open source context layer for agents and humans. Catalog every service, API, queue, topic, database and pipeline, then expose real, governed context to AI agents and your team. Self-host it free, or run it on Marmot Cloud at cloud.marmotdata.io.
This file contains links to documentation sections following the llmstxt.org standard.
## Documentation
- [Claude Agent SDK](https://marmotdata.io/docs/Agents/claude-agent): The Claude Agent SDK integration ships in the Python and TypeScript SDKs. It has two halves:
- [Marmot for Agents](https://marmotdata.io/docs/Agents/): An agent acting without context is guessing. It doesn't know which table holds customer orders, who owns the payments service, what a column means,...
- [LangChain](https://marmotdata.io/docs/Agents/langchain): The LangChain integration ships in the Python and TypeScript SDKs. It has two halves:
- [AWS](https://marmotdata.io/docs/Cloud/Guides/aws): Your first pipeline on Marmot Cloud, cataloging the Glue Data Catalog with no access key.
- [Azure](https://marmotdata.io/docs/Cloud/Guides/azure): Your first pipeline on Marmot Cloud, cataloging Blob Storage with no client secret.
- [Google Cloud](https://marmotdata.io/docs/Cloud/Guides/google): Your first pipeline on Marmot Cloud, cataloging BigQuery with no service account key.
- [Access control](https://marmotdata.io/docs/Cloud/access-control): An IAM layer for the catalog. Roles bound on the organization, a data product, an asset, a glossary term or a secret store, through a policy API wi...
- [Getting started](https://marmotdata.io/docs/Cloud/getting-started): Sign up for Marmot Cloud, launch an instance, sign in from the CLI and start a Terraform configuration.
- [Marmot Cloud](https://marmotdata.io/docs/Cloud/): Marmot, run for you. A managed instance on its own hostname, with fine-grained access control, federated secret stores and keyless pipelines.
- [Instances](https://marmotdata.io/docs/Cloud/instances): Regions, sizes, versions and the lifecycle of a Marmot Cloud instance.
- [Pipelines](https://marmotdata.io/docs/Cloud/pipelines): On Marmot Cloud a pipeline authenticates to your cloud as itself, or reads a secret from your vault at run time. No credential is stored.
- [Private plugin registry](https://marmotdata.io/docs/Cloud/plugin-registry): Push your own Marmot plugins to your instance and run them like any core plugin.
- [Secret stores](https://marmotdata.io/docs/Cloud/secret-stores): Point Marmot at Secret Manager, Secrets Manager, Key Vault or Vault through a federated identity. Marmot stores the address of a secret, never its ...
- [Security and isolation](https://marmotdata.io/docs/Cloud/security): How Marmot Cloud instances are kept apart, what an instance is allowed to reach, and how your data is protected.
- [Single sign-on](https://marmotdata.io/docs/Cloud/single-sign-on): Connect Okta, Entra ID, Auth0, Keycloak or any OIDC provider to your instance from the Marmot Cloud console, and map provider groups onto Marmot ro...
- [Auth0 OIDC](https://marmotdata.io/docs/Configure/Authentication/auth0-oidc): Configure Auth0 as an authentication provider
- [Generic OIDC](https://marmotdata.io/docs/Configure/Authentication/generic-oidc): Configure any OIDC-compliant provider for authentication
- [GitHub OAuth](https://marmotdata.io/docs/Configure/Authentication/github-oauth): Configure GitHub as an authentication provider
- [GitLab OIDC](https://marmotdata.io/docs/Configure/Authentication/gitlab-oidc): Configure GitLab as an authentication provider
- [Google OIDC](https://marmotdata.io/docs/Configure/Authentication/google-oidc): Configure Google as an authentication provider
- [Authentication](https://marmotdata.io/docs/Configure/Authentication/): Configure authentication providers
- [Keycloak OIDC](https://marmotdata.io/docs/Configure/Authentication/keycloak-oidc): Configure Keycloak as an authentication provider
- [Okta OIDC](https://marmotdata.io/docs/Configure/Authentication/okta-oidc): Configure Okta as an authentication provider
- [Slack OIDC](https://marmotdata.io/docs/Configure/Authentication/slack-oidc): Configure Slack as an authentication provider
- [Users, roles and access](https://marmotdata.io/docs/Configure/access-control): Who can sign in to open source Marmot, what each role grants, how service accounts and API keys work, and how to check what a principal can actuall...
- [Anonymous Authentication](https://marmotdata.io/docs/Configure/anonymous-access): Marmot allows you to enable anonymous authentication, which provides restricted access to the application without requiring users to log in.
- [Customisable Banner](https://marmotdata.io/docs/Configure/banner): Marmot allows you to display a customisable banner at the top of the application to communicate important information to users, such as maintenance...
- [Elasticsearch](https://marmotdata.io/docs/Configure/elasticsearch): Marmot can optionally use Elasticsearch to enhance search with deep fuzzy matching across all record fields including metadata, descriptions, docum...
- [Configure](https://marmotdata.io/docs/Configure/): Marmot is configured using a YAML file or environment variables. All settings have sensible defaults so you only need to specify what you want to c...
- [Table Preview](https://marmotdata.io/docs/Configure/table-preview): Table preview is an **experimental** feature that allows users to see a sample of rows directly in the asset detail page. Because it requires a liv...
- [Telemetry](https://marmotdata.io/docs/Configure/telemetry): Marmot collects **anonymous** telemetry to help us understand how the product is used and where to focus development effort. Telemetry is enabled b...
- [TLS](https://marmotdata.io/docs/Configure/tls): To enable TLS, provide a certificate and private key:
- [CLI / Binary](https://marmotdata.io/docs/Deploy/CLI): Run Marmot directly on your system using the single binary.
- [Docker Compose](https://marmotdata.io/docs/Deploy/Docker-Compose): Deploy Marmot and PostgreSQL together with Docker Compose.
- [Docker](https://marmotdata.io/docs/Deploy/Docker): Deploy Marmot using Docker containers with your own PostgreSQL database.
- [Helm / Kubernetes](https://marmotdata.io/docs/Deploy/Helm): Deploy Marmot to your Kubernetes cluster using our official Helm chart.
- [Deploy](https://marmotdata.io/docs/Deploy/): There are multiple ways to deploy Marmot - choose whichever method works best with your existing infrastructure and workflows. If you would rather ...
- [Creating a Marmot Plugin](https://marmotdata.io/docs/Develop/creating-plugins): This guide walks you through creating a simple HelloWorld plugin for Marmot that demonstrates the core concepts of plugin development.
- [Local Development](https://marmotdata.io/docs/Develop/local-development): 1. Start PostgreSQL:
- [Translations](https://marmotdata.io/docs/Develop/translations): The web UI reads its text from a message catalog, so a new language is a JSON file rather than a fork of the UI. English (...
- [Claude Code](https://marmotdata.io/docs/MCP/claude-code): Anthropic's CLI for Claude AI with native MCP support.
- [Claude Desktop](https://marmotdata.io/docs/MCP/claude-desktop): Anthropic's official desktop application with native MCP support.
- [Cline](https://marmotdata.io/docs/MCP/cline): VS Code extension for autonomous AI assistance with native MCP support.
- [Cursor](https://marmotdata.io/docs/MCP/cursor): AI-first code editor with native MCP support.
- [Model Context Protocol (MCP)](https://marmotdata.io/docs/MCP/): Marmot includes a built-in **Model Context Protocol (MCP)** server that enables AI assistants like Claude, ChatGPT and other LLM-powered tools to i...
- [LibreChat](https://marmotdata.io/docs/MCP/librechat): Universal AI chat interface supporting multiple providers with native MCP support.
- [Skills](https://marmotdata.io/docs/MCP/skills): Give any AI coding agent knowledge of the Marmot CLI, REST API and MCP server.
- [Notifications](https://marmotdata.io/docs/Notifications/): Stay informed about changes to your data assets with configurable notifications.
- [Webhooks](https://marmotdata.io/docs/Notifications/webhooks): Send team notifications to Slack, Discord, or any HTTP endpoint.
- [Plugins](https://marmotdata.io/docs/Plugins/): Plugins automatically discover and catalog your data assets in Marmot. They connect to external systems, extract metadata and lineage, and create a...
- [API](https://marmotdata.io/docs/Populating/API): Marmot provides a comprehensive RESTful API to programmatically interact with your data catalog. This API allows you to create, read, update, and d...
- [CLI](https://marmotdata.io/docs/Populating/CLI): The `ingest` command discovers metadata from configured data sources and catalogs them as assets in Marmot. It supports multiple data sources, can ...
- [Kubernetes Operator](https://marmotdata.io/docs/Populating/Operator): Ingest assets into your catalog declaratively using the Marmot Operator.
- [Pipelines](https://marmotdata.io/docs/Populating/Pipelines): A plugin pointed at a source, run by your Marmot server on a schedule. Declare it in Terraform or create it in the UI.
- [Pulumi](https://marmotdata.io/docs/Populating/Pulumi): Using Pulumi with the Marmot Terraform provider provides a powerful "Data Catalog as Code" approach, allowing you to define, version control, and a...
- [Terraform](https://marmotdata.io/docs/Populating/Terraform): Using Terraform, you can manage Marmot as code, declaring your Marmot resources and automating them alongside the rest of your infrastructure.
- [UI](https://marmotdata.io/docs/Populating/UI): The Runs page creates and manages pipelines from a form. It is the quickest way to try a plugin. A pipeline people rely on belongs in Terraform, so...
- [Populating your catalog](https://marmotdata.io/docs/Populating/): There are two ways to fill the catalog. A pipeline points a plugin at a source and discovers what is there on a schedule. Declaring assets as code ...
- [SDK](https://marmotdata.io/docs/SDK/): The Marmot SDK is a typed client for the REST API, available in **Python**, **Go** and **TypeScript**. Authentication resolves automatically from e...
- [API Reference](https://marmotdata.io/docs/api-reference): Marmot exposes a REST API for managing assets, lineage, glossary terms, data products, and more.
- [Asset Rules](https://marmotdata.io/docs/asset-rules): Asset Rules automatically apply enrichments to assets matching specific criteria. Define a rule once and Marmot keeps everything in sync as your Ca...
- [CLI Reference](https://marmotdata.io/docs/cli): The Marmot CLI lets you interact with your data catalog directly from the terminal.
- [Data Products](https://marmotdata.io/docs/data-products): Data Products let you group related assets into logical collections. A "Customer Analytics" product might bundle together a PostgreSQL table storin...
- [Glossary](https://marmotdata.io/docs/glossary): The Glossary lets you define business terms and create a shared vocabulary across your organisation. Instead of different teams using different nam...
- [Introduction](https://marmotdata.io/docs/introduction): Who builds Marmot, what it is for, what we stand for, and where to go next in the documentation.
- [Metrics](https://marmotdata.io/docs/metrics): Marmot collects various application metrics for both Prometheus monitoring and built-in dashboards in the UI.
- [OpenLineage](https://marmotdata.io/docs/open-lineage): OpenLineage is an open standard for data lineage collection and analysis. It provides a unified way to track data flows across different tools and ...
- [Query Language](https://marmotdata.io/docs/queries): Marmot provides a query language for searching and filtering assets in your catalog. The language supports free-text search, field-specific filters...
- [Quick Start](https://marmotdata.io/docs/quick-start): Get Marmot running in seconds with Docker Compose.
## Resources
- [What Is an AI Context Layer? A 2026 Guide](https://marmotdata.io/resources/ai-context-layer): What an AI context layer is, why AI agents need one, how it differs from RAG, and how to build one on a data catalog with MCP. A plain, vendor-neut...
- [AI Data Engineering: Giving Agents Real Context](https://marmotdata.io/resources/ai-data-engineering): What AI data engineering is, why context matters more than the model, and how to give LLMs and agents real, governed context about your data so the...
- [What Is a Data Catalog? A 2026 Guide](https://marmotdata.io/resources/data-catalog): What a data catalog is, what it does, its features and benefits, and how to choose one in 2026 when AI agents query metadata at runtime. A plain, v...
- [Best Data Catalogs for AI Agents in 2026: The AI Context Layer Compared](https://marmotdata.io/resources/data-catalogs-for-ai-agents): The best data catalogs for AI agents in 2026, compared on native MCP, governed context and footprint. Marmot, OpenMetadata, DataHub, Atlan, Collibr...
- [Data Governance for the AI Era: A 2026 Guide](https://marmotdata.io/resources/data-governance): What data governance is, what it covers, best practices, and how it works when AI agents act on your data: ownership, policies and access control e...
- [Data Quality for AI Agents: A 2026 Guide](https://marmotdata.io/resources/data-quality): What data quality is, its dimensions, the tools that measure it, and why freshness and certification matter more when an AI agent acts on metadata ...
- [Marmot Resources: Data Catalogs, Governance and AI Context](https://marmotdata.io/resources/): Guides and comparisons on data catalogs, data governance, data quality, AI data engineering and MCP for data. Practical, vendor-neutral resources f...
- [Marmot vs Atlan: AI Context Layer Comparison (2026)](https://marmotdata.io/resources/marmot-vs-atlan): Marmot vs Atlan compared as AI context layers for agents. An open source catalog you self-host as a single Go binary versus Atlan's fully managed, ...
- [Marmot vs DataHub: AI Context Layer Comparison (2026)](https://marmotdata.io/resources/marmot-vs-datahub): Marmot vs DataHub compared as AI context layers for agents. Native built-in MCP on a single binary versus DataHub's official MCP server on a Kafka,...
- [Marmot vs OpenMetadata: AI Context Layer Comparison (2026)](https://marmotdata.io/resources/marmot-vs-openmetadata): Marmot vs OpenMetadata compared as AI context layers for agents. Native MCP on a single Go binary and Postgres versus OpenMetadata's broad connecto...
- [MCP for Data: Connecting AI Agents to Your Catalog](https://marmotdata.io/resources/mcp-for-data): What the Model Context Protocol (MCP) is, how a data catalog uses it to expose governed context to AI tools like Claude and Cursor, and how it comp...
## Blog
- [Marmot: Data catalog without the complex infrastructure](https://marmotdata.io/blog/data-catalog-without-complex-infrastructure): Data catalogs shouldn't need a platform team to run. Meet Marmot, an open source data catalog that needs only PostgreSQL, ships as a single Go bina...
- [Postgres: One Database to Rule Them All](https://marmotdata.io/blog/postgres-one-database-to-rule-them-all): Why PostgreSQL alone can back a production data catalog. How Marmot replaces Kafka, Elasticsearch and a search cluster with one Postgres database, ...
- [AI is making assumptions about your data, and getting them wrong](https://marmotdata.io/blog/ai-is-making-assumptions-about-your-data): AI assistants write code well but are blind to your data: what tables exist, who owns them, what business terms mean. Why agents need a governed so...
- [Deploy Marmot to Google Cloud Run](https://marmotdata.io/blog/deploy-marmot-on-google-cloud-run): A step-by-step guide to deploying Marmot on Google Cloud Run with Cloud SQL for PostgreSQL, defined in Terraform. A serverless data catalog that sc...
- [Connect Claude Desktop to Marmot](https://marmotdata.io/blog/connect-marmot-to-claude-desktop): Half the company already has Claude Desktop installed, and the knowledge about your data is scattered across tools, wikis and people's heads. Conne...
- [Catalog your Kubernetes clusters](https://marmotdata.io/blog/kubernetes-context-in-marmot): Marmot's new Kubernetes, Amazon Elastic Kubernetes Service and Google Kubernetes Engine plugins catalog your clusters, so services, deployments and...
- [Migrate from OpenMetadata to Marmot in five minutes](https://marmotdata.io/blog/import-openmetadata-into-marmot): Marmot's new OpenMetadata plugin imports an entire OpenMetadata instance in one run: tables, topics, dashboards, pipelines, lineage and the busines...
