tura

Tura is an open-source agent runtime harness. According to the project, it delivers better results with fewer tokens by turning a task into a runtime-managed command graph, so deterministic execution continues without another model round trip. The project states that Tura executes in one turn the same workflow that a ReAct architecture needs five turns to complete.

Tura exposes one macro tool, command_run, instead of dozens of small tools. The agent can build a multi-step execution tree and run related actions in one LLM turn. In published DeepSWE comparisons, Tura Direct used 77.5% fewer aggregate tokens than Codex CLI with a comparable verifier success rate, while Tura Balanced reached an 80.0% success rate and used 31.1% fewer tokens.

Tura treats context as part of the runtime state machine. It uses task_status, runtime prompts, and recursive execution manuals to keep active context scoped to the current task. Because compaction is a CLI operation, Tura can preserve exact execution state in task_status.compact_context. Tura does not bundle provider credentials; on first launch, users configure an LLM provider and select a model before sending a prompt.