CC Switch AI Control Plane

Overview

CC Switch is a local control plane for coordinating providers, heterogeneous AI clients, native configuration, request routing, workflow context, evidence, and recovery. The series follows provider choices from durable database records through client-specific projections, then into an optional local proxy that can translate requests, log routed traffic, and apply bounded failover. It distinguishes MCP servers, prompts, skills, sessions, and OpenClaw workspace files by their ownership and activation rules, and treats dashboards, imports, sync, OAuth, and network listeners as evidence and trust boundaries rather than guarantees. The central lesson is proportionality: use direct projection when simplicity is enough, add takeover only for request-plane needs, and verify the actual client, model, tools, identity, and version at the boundary that matters.

Learning outcomes

CC Switch coordinates provider intent with client-specific projections; it does not make those layers one source of truth.
Takeover enables request-time routing, translation, logging, and failover, but adds proxy lifecycle and trust boundaries.
Failover can reduce interruption only when queued providers fit the task's model, tool, timing, and identity requirements.
Portable context, usage data, backups, and sync remain bounded by app formats, evidence coverage, version drift, and input trust.

Keywords

CC SwitchprovidersconfigurationproxyroutingfailoverMCPpromptsskillsobservability

Chapters

CC Switch: AI Tooling Fragmentation

Keeping provider state, client files, and requests coherent

9 min · Keeping provider state, client files, and requests coherent

9 min

CC Switch Provider State Model

How cards, live files, and processes reconcile

11 min · How cards, live files, and processes reconcile

11 min

CC Switch Request Routing

How takeover moves provider choice into the request path

12 min · How takeover moves provider choice into the request path

12 min

CC Switch Failover Policy

Queues, circuit breakers, and limits on recovery

12 min · Queues, circuit breakers, and limits on recovery

12 min

CC Switch Context Portability

App-specific ownership for MCP and workflow files

13 min · App-specific ownership for MCP and workflow files

13 min

CC Switch Observability and Recovery

Reading logs, estimates, backups, and trust boundaries

14 min · Reading logs, estimates, backups, and trust boundaries

14 min

Choosing the Right Boundary for CC Switch

When direct mode, takeover, failover, or portability fits

12 min · When direct mode, takeover, failover, or portability fits

12 min

Preview every chapter · Unlock full text, audio, and video from chapter 4

Plans

Ch.1-3 FreeCh.4-7 Premium

FAQ

What does CC Switch actually manage?

CC Switch manages provider records and projections, optional request routing, workflow context, usage evidence, and recovery state; each client still owns its native runtime behavior.

When should CC Switch use a local proxy?

Use a local proxy when hot switching, protocol conversion, routed logs, or bounded failover justify a request-plane service; direct mode is simpler when native compatibility and occasional switching are enough.

Does CC Switch failover preserve the same workflow?

Not necessarily. Failover can preserve an attempt at availability, but the backup may change the model, tool behavior, latency, price, or account identity.

How reliable are CC Switch usage and health indicators?

They are local evidence, not universal guarantees: proxy logs cover routed traffic, session imports cover parseable client history, and cost or health values may not match provider billing or future behavior.