Databricks Model Serving

databricks/databricks-agent-skillsOfficial

This gives Claude the commands to create, configure, and query Databricks Model Serving endpoints without you needing to remember CLI flags. It handles the full lifecycle: spinning up pay-per-token foundation model APIs, provisioned throughput endpoints with dedicated GPUs, or custom MLflow models. The workflow is sensible: it makes you discover commands with help flags first instead of guessing syntax, walks through polling for readiness states, and shows how to wire endpoints into Databricks Apps with the right YAML incantations. Most useful when you're toggling between endpoint types or debugging why something is stuck in NOT_READY. The OpenAPI schema command is a nice touch for figuring out what payload format a custom model actually expects.

skills.sh installs: 756
GitHub stars: 393
last updated: Aug 21, 2026

View on GitHub

Install

$ npx -y skills add databricks/databricks-agent-skills --skill databricks-model-serving --agent claude-code

Installs into .claude/skills of the current project.

Files

Model Serving Endpoints

FIRST: Use the parent databricks-core skill for CLI basics, authentication, and profile selection.

Gateway boundary: This skill owns legacy AI Gateway configuration attached to a Model Serving endpoint, such as serving-endpoints put-ai-gateway. An explicit Unity Gateway / Unity AI Gateway request or any databricks ai-gateway model service, MCP service, or model provider service belongs to databricks-unity-gateway.

Model Serving provides managed endpoints for serving LLMs, custom ML models, and external models as scalable REST APIs. Endpoints are identified by name (unique per workspace).

Endpoint Types

Type	When to Use	Key Detail
Pay-per-token	Foundation Model APIs (Llama, GPT-5, Claude, Gemini, etc.)	Uses `system.ai.*` catalog models, pre-provisioned in every workspace. Discover at runtime — see Foundation Model API endpoints below.
Provisioned throughput	Dedicated GPU capacity	Guaranteed throughput, higher cost
Custom model	Your own MLflow models or containers	Deploy any model with an MLflow signature