Clarify LLM Provider Terminology Across the Platform

[a5anka]

Problem

The current terminology around LLM providers is confusing for users. Specifically:

• "LLM Service Providers" vs "LLM Providers" — the distinction between these two terms is unclear. Users struggle to understand what each refers to and how they relate.
• As new features reference LLM providers, the naming inconsistency will create more confusion across the platform.

Current Flow

1. Org-level (Admin persona): Configure LLM providers (e.g., OpenAI, Anthropic) with connection details, security settings, rate limiting, and org-wide guardrails.
2. Agent-level (Developer persona): Use a configured LLM provider and apply agent-specific guardrails.

The terminology should make this two-level flow intuitive.

Suggestion

1. Standardize on "LLM Providers" as the single term across the platform. Drop "LLM Service Providers" entirely.
2. The same term is used at both levels - the difference is the action: configure at org level, use at agent level. Guardrails are managed at both levels independently.

Related

• [Design Proposal] LLM Provider management #286 — [Design Proposal] LLM Provider management
• [Design Proposal] Consume existing LLM Providers from an Agent #385 — [Design Proposal] Consume existing LLM Providers from an Agent