Your agents are capable. Their separate terminals leave you managing the project between them.
Different supported agents contribute to the same task workflow.
Start with the agent you already use
Connect a runner on a machine with the supported CLI installed and authenticated. The runner provides the execution environment; Norc provides the project, task assignment, and context around the run.
Your provider account, subscription or API billing, local tools, and repository access remain part of that environment. Adding Norc does not replace those requirements or bundle agent usage.
Split the workflow by responsibility
For example, one agent can investigate the current export behavior, another can implement a scoped change, and a follow-up task can update the documentation. Assign the roles deliberately and connect the dependencies.
Norc can carry finished dependency outputs into later tasks. The coding workflow can use isolated Git worktrees, with run evidence and branch or pull-request information available for review. Keep merge and release decisions aligned with your team’s rules.
Keep shared context without assuming identical behavior
Claude Code and Codex do not have identical runtime behavior. Norc coordinates their supported execution paths; it does not make their models, native tools, or permission systems interchangeable.
Check the skills, credentials, repository state, and allowed actions for each agent. Run a small task first and inspect its result before assigning a larger queue.
Practical questions.
Do I need both Claude Code and Codex?
No. Start with one supported agent. Add another when a workflow benefits from a second role or execution option.
Is agent usage included?
Your agent subscriptions or API charges and runner infrastructure remain separate. Norc coordinates those agents and records supported run information.
By the Norc team · Product details checked against the Norc project documentation. View the project