Agent orchestration runtime with long-term memory.
A higher-level orchestration runtime (the NanoClaw / OpenClaw pattern) that gives agents durable, long-term memory across runs.
Where agent is one loop, claw coordinates agents and the state that outlives a
single conversation.
Highlights
- Long-term memory that persists across agent runs
- Orchestration layer above the single-agent loop
- Configurable runtime for multi-agent patterns
Try it
uv run xdog --help