Choose from supported cloud and self-hosted models in one catalog. Manage access and credentials centrally, so teams can use the right approved model for each task.

Connect agents to business tools and data through Model Context Protocol (MCP). Discover available tools and control which agents can access and use them.

Manage agents, versions and deployments in one place. Track changes and move tested versions between environments to support a controlled path from development to production.

Define who can use which models, tools and platform resources. Enforce access policies as agents work, with approval workflows for sensitive platform changes.

Follow agent activity, investigate failures and review model and tool usage. Execution traces and audit records help teams understand what happened and support operational oversight.

Help agents retain and retrieve relevant information for longer tasks and ongoing interactions. Configure what persists between tasks and how memories are shared.

Set budgets and usage limits for teams and agents. Control model usage, tool calls and task duration to help manage spend and keep automated work within defined limits.

Compare agent versions, identify bottlenecks and evaluate results against your chosen criteria. Use tuning workflows to improve prompts and settings, balancing quality, speed and cost.