Runtive

Stop re-explaining your codebase to your coding agent

Every time you start Claude Code or Codex, the agent knows nothing about the last session — what was decided, which constraints were set, what was left unfinished. Runtive closes that loop. One command at the end of a task captures the session, diff and checks, grades the work, and refreshes the context your next run starts from.

Pricing

Runtive Workspace — $19 USD / month, with a 14-day free trial. Billed monthly, cancel anytime. See what is included.

How it works

• Finish a task with runtive finish.
• Runtive captures the session transcript, the git diff and your check output.
• You get a Decision Brief — Accept, Review or Risky — grounded in that

evidence, plus a structured extract of decisions, current state and the recommended next step.

• Your next Claude Code or Codex session reads the refreshed context

automatically.

Inference is customer-funded: Runtive uses your own Claude Code or Codex access, or your own API key, and never holds your model credentials.

Contact

Support, security and privacy: support@runtive.dev