---
title: "Application Profiles vs Per-Client Config: Where OAuth Policy Lives"
description: "Application Profiles bundle token, rotation, and DPoP policy once and attach it to many OAuth clients. This comparison shows when a shared profile beats configuring each client by hand."
---

# Application Profiles vs Per-Client Config: Where OAuth Policy Lives

**Type:** [Comparison](https://yomu.fyi/visual-explainers/category/comparison.md)

![A two-column comparison diagram: on the left, per-client OAuth configuration with settings repeated on every client card; on the right, one Application Profile connected to several client chips. A bottom bar gives the decision cue based on client count and churn.](https://yomufyi.s3.us-east-1.amazonaws.com/visual-explainers/532e37eb-5718-48fc-8a06-3c059c04af41/image-b7619c23-5bf6-4fd3-bf7d-b4ba1adfc256.webp)

> Application Profiles bundle token, rotation, and DPoP policy once and attach it to many OAuth clients. This comparison shows when a shared profile beats configuring each client by hand.

## Explanation

Application Profiles are a reusable bundle of security policy — token lifetimes, refresh token rotation, DPoP binding, and allowed grant types — that an identity provider applies to many OAuth clients at once. Instead of hand-configuring each client, you define the profile once and attach it as clients arrive, whether pre-registered by an admin or discovered just-in-time during an authorization flow. The trade-off is precision versus scale: per-client settings are exact but drift and slow onboarding, while a shared profile scales but forces you to decide what belongs in the default. That matters because agent and MCP clients multiply faster than teams can review them one by one.

## FAQ

### What actually goes inside an Application Profile?

A profile holds the security policy an identity provider applies to a group of OAuth clients, such as refresh token rotation, DPoP binding, and allowed authentication methods. In Auth0's implementation, profiles let dynamically registered clients inherit consistent policy instead of being configured one at a time. Treat it as a policy bundle, not a client record.

### Manual pre-registration or dynamic registration — which should I pick?

Manual pre-registration has an administrator submit a client's metadata URL for review and approval, which suits a small number of known partners. Dynamic registration discovers clients just-in-time during an authorization flow, which suits many agent or MCP clients that appear without warning. The registration mode decides how a client arrives; the profile decides what it may do.

### What breaks if every client gets its own settings?

Per-client configuration is precise, but the settings drift apart over time and every new client becomes a manual review. That cost grows quickly when agent and MCP clients arrive faster than a team can review them, and it makes policy changes hard to roll out consistently. A shared profile trades some per-client precision for one place to change and audit the default.

**Tags:** [Application Profiles](https://yomu.fyi/topic/application-profiles.md), [Authentication](https://yomu.fyi/topic/authentication.md), [Authorization](https://yomu.fyi/topic/authorization.md), [Governance](https://yomu.fyi/topic/governance.md), [MCP](https://yomu.fyi/topic/mcp.md)

## Related reading

### [Air Governance: Manage AI Agents, Access, and Spend Throughout Your Organization](https://yomu.fyi/post/air-governance-manage-ai-agents-access-and-spend-throughout-your-organ.md)
- Company: [Jetbrains](https://yomu.fyi/company/jetbrains.md)
- Author: Viktor Kiselev
- Published: Oct 22, 2026
- Type: Announcement

JetBrains Air Governance, formerly JetBrains Central, is now generally available as the organization layer of JetBrains Air for managing agentic AI usage. It provides a single console to manage AI agents, models, and providers across teams, set spending limits, and track both costs and outcome metrics like the split between human-authored and AI-generated code. The tool works with Claude Agent, Codex, Gemini CLI, Junie, and other ACP-compatible agents across IDEs, CLI, web, and mobile, and supports multiple vendors without separate contracts. Admins can set AI credit limits per user, team, or service account, manage BYOK providers, approve agents, and configure MCP servers at the company level, combining these into policies for different groups. Air Governance is available to all JetBrains organizational customers with free and paid plans starting at USD 20 per developer per month, and it already manages usage for roughly 2,300 people at JetBrains.


### [Agentic partners integration: AI support agents for Microsoft Azure Red Hat OpenShift](https://yomu.fyi/post/agentic-partners-integration-ai-support-agents-for-microsoft-azure-red.md)
- Company: [Red Hat](https://yomu.fyi/company/red-hat.md)
- Author: Sharon Dashet, Carlos Camacho Gonzalez
- Published: Oct 22, 2026
- Type: Explainer

The operations team troubleshooting production issues on Azure Red Hat OpenShift must already know which tools to reach for, switching between portals, CLI, and dashboards, with tickets arriving incomplete and triage taking hours. The agentic-partners-integration quickstart solves this by connecting a central orchestration layer with domain-specific agents, including a Microsoft Azure MCP Server that exposes Azure resource management operations as MCP tools and an agent that discovers them at runtime. In an OOMKilled investigation, the agent correlates resource health, CPU metrics, and recommendations using a discover-then-execute pattern to produce a data-backed diagnosis. Security is layered: the service principal defaults to Azure Reader role, Open Policy Agent policies gate every request, and each tool call is logged for audit. This pattern extends beyond Azure; any infrastructure backend exposing an MCP server can become an agent capability without framework changes.


### [Give your Supabase app its own MCP server](https://yomu.fyi/post/give-your-supabase-app-its-own-mcp-server.md)
- Company: [Supabase](https://yomu.fyi/company/supabase.md)
- Author: raul\_barroso,gregnr,saxon\_fletcher
- Published: Oct 22, 2026
- Type: Tutorial

Supabase announced the MCP Server block that lets developers give their Supabase apps an MCP server, enabling customers to use app actions as tools from clients like ChatGPT or Claude after signing in with their existing account. Every tool call runs as that user, so RLS policies decide what each agent can access. The block adds an Edge Function using the official MCP TypeScript SDK and Supabase authentication helpers to verify tokens and provide a user-specific Supabase client. Developers write tool functions that can query RLS-protected tables, call database functions, or invoke Edge Functions, keeping business rules in the code the tool calls. The server can also display app UI inside the chat via MCP Apps, and the block includes a Copy agent prompt button and works with Cloudflare Workers and other Web-standard backends.


### [Scaling AI Agent Identity with CIMD and Application Profiles in Auth0](https://yomu.fyi/post/scaling-ai-agent-identity-with-cimd-and-application-profiles-in-auth0.md)
- Company: [Auth0](https://yomu.fyi/company/auth0.md)
- Author: Eric Wu
- Published: Oct 22, 2026
- Type: Announcement

The rise of AI agents and MCP-based ecosystems introduces identity challenges when onboarding numerous clients without manual overhead. Auth0 expands its Client ID Metadata Document (CIMD) support with two registration modes — manual pre-registration and dynamic registration — for OAuth clients. Manual pre-registration requires an administrator to submit a metadata URL for review and approval, while dynamic registration enables just-in-time client discovery during authorization flows. CIMD uses an HTTPS URL as the client identifier pointing to a metadata document with the client’s name, redirect URIs, and authentication methods; Auth0’s implementation uses shared caching to reduce database write pressure and support more dynamic clients without latency. Application Profiles centrally manage security policies such as refresh token rotation and DPoP across dynamically registered clients, aiming to provide a practical identity foundation for MCP servers and agentic applications.


### [Introducing Terraform Auth0 Modules](https://yomu.fyi/post/introducing-terraform-auth0-modules.md)
- Company: [Auth0](https://yomu.fyi/company/auth0.md)
- Author: Rajat Bajaj
- Published: Oct 22, 2026
- Type: Announcement

Teams building B2B SaaS on Auth0 repeatedly configure applications, APIs, M2M clients, MFA, attack protection, branding, custom domains, and enterprise SSO. The new Terraform Auth0 Modules package this foundation into ready-made modules, starting with the b2b-saas-starter module on the Terraform Registry. A single module block creates an application, API, tenant settings, attack protection, and MFA, with security defaults like RBAC enforcement, PKCE, rotating refresh tokens, and TOTP MFA. Roles, permissions, and enterprise SSO connections for customer organizations are described as data inside the module, which validates input to prevent errors. The module supports multiple applications per tenant by letting other instances skip tenant-wide resources, and each default can be overridden via module inputs.


### [Survey Says: You'll swap the model, but everything you build around it is staying put](https://yomu.fyi/post/survey-says-you-ll-swap-the-model-but-everything-you-build-around-it-i.md)
- Company: [Stack Overflow](https://yomu.fyi/company/stack-overflow.md)
- Author: Craig McLuckie
- Published: Oct 21, 2026
- Type: Explainer

A Stacklok survey of 520 technical leaders at North American enterprises running LLMs or agents found that only 8% use a single model provider, while 25% use five or more. The hardest elements to migrate are integrations to internal systems or MCP servers (53%) and multi-step agent workflows (52%). Although 84% say avoiding lock-in is important, “ability to switch model providers without rebuilding” ranked sixth out of eight criteria, and “avoiding dependence on a single vendor” ranked last. Organizations using five or more model providers are more likely to build integrations (76% vs 42%), use open source agent frameworks (45% vs 22%), and prioritize portability (67% vs 47%). The survey also found that 80% would evaluate an open alternative to closed frontier architectures, and 88% expect to take more control of their AI architecture as they gain experience.
