hillclimb run
Start a run on a problem or a run-spec YAML.
Start a run on a problem or a run-spec YAML.
The search runs as a detached engine and the terminal comes straight back: hillclimb watch follows it, hillclimb stop --all ends it. --no-detach keeps it in this terminal instead (Ctrl-C stops it).
hillclimb run [OPTIONS] TARGETArguments
| argument | what it is |
|---|---|
TARGET | — required |
Options
| option | what it does |
|---|---|
--budget TEXT | Wall-clock budget, e.g. 2h / 30m. |
--agent, --backend TEXT | The coding agent that runs the operators: claude-code | codex | pi | dummy (--backend is the old spelling). |
--model TEXT | Model the agent runs, e.g. sonnet / opus. |
--climber TEXT | The climber (default: greedy): a bundled name, a directory holding climber.yaml, or one .py file; params via config climber.params. Repeat it (--climber greedy --climber gepa) for a mixed fleet: one search per climber on the problem, under one run, each tagged as an arm. |
--holdout / --no-holdout | Hidden selection holdout. Default: on. |
--learning / --no-learning | Cross-search memory (cards/claims injection + distillation); off = memory-blind arm. Default: on. |
--name TEXT | Run name shown in the TUI. |
--parallel-agents, --parallel-operators INTEGER | Concurrent agents per search, one candidate each (--parallel-operators is the old spelling). |
--parallel-searches INTEGER RANGE | Independent searches on the problem at once, each its own detached engine. Default: 1. |
--detach / --no-detach | Run as a detached engine (the default; hillclimb watch follows it) or in this terminal (Ctrl-C stops it). Default: on. |
--n-replicates, --n-trials INTEGER | Seeded runs per trial (the median is the trial's score; --n-trials is the old spelling). |
--seed-from PATH | Incumbent solution.py scored as the floor candidate. |
--set TEXT | Any config setting, dotted: --set climber.ref=openevolve --set learning.enabled=false. |
--arm-set TEXT | A setting for one arm of a mixed fleet, ARM:KEY=VALUE: --arm-set gepa:concurrency.parallel_agents=1 (applied after --set). |
--experiment TEXT | Tag the search as one arm of an experiment (with --arm); names a mixed fleet's experiment. |
--arm TEXT | The arm name (with --experiment). |