Skip to main content

AI agents are only as good as the context they can reach.

Marmot is an open source context layer for engineers and AI agents. It tracks schemas, ownership and lineage across your entire stack.

Get startedLive demo

Manage via

Query via

CLI
Terraform
Pulumi
REST API
SDK
Marmot
  • Assets
  • Lineage
  • Ownership
  • Glossary
MCP server
Web UI
REST API
CLI

Manage via

CLI
Terraform
Pulumi
REST API
SDK

Databases

Warehouses & lakes

Streaming & pipelines

Compute

Marmot
  • Assets
  • Lineage
  • Ownership
  • Glossary

Query via

MCP server
Web UI
REST API
CLI

via MCP

Everyone answers their own questions

Where a number came from. Whether a column is safe to change. Who owns this table. Questions that used to sit in the data team's queue. Marmot answers them without anyone needing to know where to look.

Same goes for agents. They are limited to whatever context you paste into the prompt. With Marmot they can go and find it themselves.

Set up MCP

Where do we hold customer PII?

discover_data

Four assets are tagged pii: the customers, orders and support_tickets tables in Postgres, plus the user_events topic in Kafka. All four are owned by the Data Platform team.

What breaks if we rename the order_gmv column?

trace_lineage

The daily_orders_summary view and the Revenue Overview dashboard both depend on that column. The table is owned by the Data Platform team, so check with Sarah Chen first.

Up in minutes

Marmot is one service and a Postgres database you already run. Most catalog projects spend their first quarter standing up infrastructure. Marmot spends it on your data.

Quick start
~/marmot
$docker compose up -d
[+] Container postgres-1 Started
[+] Container marmot-1 Started
Marmot is running at http://localhost:8080

Load tested with a single Postgres and a single Marmot process:

500k+Assets
100+Concurrent users
<50msAvg response time
How we load tested it

Plugins for what you already run

Covering the databases, warehouses, lakes, streams and orchestrators your team already uses, and counting.

PostgreSQLMySQLMongoDBClickHouseDynamoDBBigQueryDuckDBTrinoIcebergDelta LakeS3Google Cloud StorageAzure Blob StorageKafkaRedpandaNATSConfluent CloudAirflowdbtLambdaKubernetesEKSElasticsearchOpenSearch

View all plugins or ask for one that isn't here yet.

See how Marmot fits into your stack.

Try the live demo, or deploy it yourself in a few minutes.

Live demoTalk to us

Built by engineers from HashiCorp, Adidas, Just Eat Takeaway.com and Traefik, who help maintain Kubernetes, Terraform, Redpanda and the Cloud Native Computing Foundation.