One command. Multiple models. No juggling.
Baya is a zero-config command-line orchestrator for AI coding agents. It turns a plain-text task list into a dependency graph and routes each task to the provider and model that fit it. Independent tasks run in parallel, dependent tasks wait for their prerequisites, and the run carries through to a report.
$ baya tasks.mdRun order · 6 tasks · 3 stages ✓ design-api codex 6 endpoints ✓ gen-schema claude 4 tables created ✓ build-ui codex orders table ✓ tests codex 14 tests pass ✓ Run complete 6 succeeded · 1m12s · 2 processes 214k tokens (188k cached) · $0.37 Flagged ⚑ gen-schema locks users ~30s at 1M+ rows $ What you get
01 LLM-planned dependency graph
Baya turns your plain-text list into a DAG. Independent tasks run in parallel; dependent tasks wait for their prerequisites.
02 Per-task model routing
Every task goes to the provider and model that fit it — codex, claude, copilot, or opencode — on the CLI subscriptions you already have. No API keys.
03 Process reuse & context handoff
Compatible tasks share one agent process, and useful findings pass forward between tasks — even across models — so you never pay twice for setup and discovery.
See all features →Multi-model AI consensus →