Skip to content

Coding Plans

Coding plans are managed as backends in the console. They are useful for coding agents, developer tools, and high-volume coding workloads where capacity is tied to a subscription or quota window.

Backends with coding plans

How Coding Plans Fit In

  1. Add the coding-plan account as a backend.
  2. Put the plan backend into a model group such as coding-agent.
  3. Use routing to define when traffic should move from the plan to a metered fallback.
  4. Monitor quota, concurrency, and fallback activity from the group detail page.

Coding group detail

Supported Plan Use Cases

  • Coding assistants and repository agents.
  • Long-running refactor or test-generation workflows.
  • Internal developer copilots.
  • Team-level plan capacity shared across products.

Put coding-plan capacity first, then add metered backends and a general-purpose fallback group. This keeps plan usage high while protecting users from quota exhaustion.

Operational Checks

  • Watch quota reset windows.
  • Track concurrency so agents do not overload a plan account.
  • Review fallback markers after heavy coding sessions.
  • Keep a metered backend ready for urgent work.

Unified AI Gateway documentation for multi-model, multi-provider, and coding-plan use cases.