tune

To optimize TeaRAGs performance, you can run hardware benchmarking to stress test embedding throughput and Qdrant storage latency, which automatically generates tuned concurrency and batching configuration parameters for your specific setup.

Indexing is slow because default concurrency and batching values may not match your hardware. Running a hardware benchmark calibrates your local embedding provider and Qdrant daemon to determine optimal throughput settings, preventing slow indexing.

Hardware benchmarking works by executing stress tests on embedding throughput and storage latency to calibrate system-specific concurrency and batching values. This process generates a tuned environment configuration for seamless MCP integration.

Benchmarking embedding throughput requires a local Qdrant daemon and the tea-rags CLI to be accessible. If local hardware resources are heavily constrained, the generated concurrency and batching values will reflect those hardware limitations in the tuned configuration.

You can tune Qdrant storage latency without manual configuration by executing automated stress tests. This benchmarks your local hardware and saves the optimal performance parameters directly to your setup progress file for immediate use.