Run AI tasks through the low-latency fast client path.
This command is optimized for hot-loop invocation and prefers fai/ai-taskd-client over full f task startup.
f fast ai:flow/noop
f fast ai:flow/bench-cli -- --iterations 30
f fast --root ~/code/flow ai:flow/dev-check
f fast --no-cache ai:flow/dev-check- Tries fast client dispatch (
fai, localtarget/.../ai-taskd-client, orai-taskd-clienton PATH). - If daemon is not running, starts
ai-taskdand retries. - Falls back to direct daemon dispatch if no fast client binary is found.
--root <PATH>: root directory for.ai/tasksdiscovery (default.)--no-cache: disable cached binary execution and use direct Moon runTASK: required AI selector likeai:flow/dev-check- trailing args after
--are passed to the task
f fastis intentionally forai:*selectors only.- Install
faifor best latency:
f install-ai-fast-client
f tasks daemon start
f fast ai:flow/noopFor pooled burst execution and timings, use fai directly:
fai --timings ai:flow/noop
printf 'ai:flow/noop\nai:flow/noop\n' | fai --batch-stdin --timings