Databricks MCP Server

Databricks MCP server enabling AI agents to interact with the Databricks Lakehouse Platform — running notebooks, executing SQL on Unity Catalog, managing clusters, accessing Delta Lake tables, running ML experiments, and orchestrating data workflows.

Evaluated Mar 06, 2026 (0d ago) vcurrent
Homepage ↗ Repo ↗ Other databricks spark delta-lake mlops mcp-server lakehouse unity-catalog
⚙ Agent Friendliness
79
/ 100
Can an agent use this?
🔒 Security
90
/ 100
Is it safe for agents?
⚡ Reliability
84
/ 100
Does it work consistently?

Score Breakdown

⚙ Agent Friendliness

MCP Quality
78
Documentation
85
Error Messages
80
Auth Simplicity
72
Rate Limits
75

🔒 Security

TLS Enforcement
100
Auth Strength
90
Scope Granularity
88
Dep. Hygiene
85
Secret Handling
88

HTTPS enforced. Unity Catalog RBAC with fine-grained access. FedRAMP, SOC 2, ISO 27001, HIPAA, PCI DSS. Enterprise Databricks security.

⚡ Reliability

Uptime/SLA
90
Version Stability
85
Breaking Changes
82
Error Recovery
80
AF Security Reliability

Best When

An agent needs to interact with a Databricks Lakehouse environment — especially for SQL analytics, ML workflows, or large-scale data processing.

Avoid When

You're using Snowflake, BigQuery, or another data platform.

Use Cases

  • Running Databricks SQL queries against Unity Catalog from analytics agents
  • Executing and monitoring Databricks notebooks from data pipeline agents
  • Accessing Delta Lake table metadata and schema from data discovery agents
  • Managing cluster lifecycle from infrastructure cost optimization agents
  • Tracking MLflow experiments and model registry from MLOps agents
  • Orchestrating Databricks Jobs from workflow automation agents

Not For

  • Teams using Snowflake, BigQuery, or Redshift as their primary platform
  • Simple SQL databases — Databricks is for large-scale data/ML workloads
  • Non-Spark workloads that don't need distributed compute

Interface

REST API
Yes
GraphQL
No
gRPC
No
MCP Server
Yes
SDK
Yes
Webhooks
Yes

Authentication

Methods: personal_access_token oauth2 service_principal
OAuth: Yes Scopes: Yes

Databricks personal access tokens, OAuth M2M for service principals, and Unity Catalog-level permissions. Service principals recommended for agents.

Pricing

Model: usage-based
Free tier: Yes
Requires CC: Yes

Databricks Unit (DBU) consumption pricing. Costs can be significant for large clusters. Community Edition free but limited. MCP server varies by implementation.

Agent Metadata

Pagination
cursor
Idempotent
Partial
Retry Guidance
Documented

Known Gotchas

  • Cluster startup time is significant (2-5 min for all-purpose, faster for SQL warehouses)
  • Workspace URL varies per region — agents need correct workspace host
  • Unity Catalog vs legacy Hive Metastore — use Unity Catalog for new workloads
  • Notebook execution is async — agents must poll for job run completion
  • DBU costs can be large for unoptimized queries — agents must be cost-aware
  • Service principal setup requires Databricks admin — not self-service

Alternatives

Full Evaluation Report

Detailed scoring breakdown, competitive positioning, security analysis, and improvement recommendations for Databricks MCP Server.

$99

Scores are editorial opinions as of 2026-03-06.

5178
Packages Evaluated
26151
Need Evaluation
173
Need Re-evaluation
Community Powered