Mikei386 e9b647e579 Add automatic per-request MCP selection

2026-08-23 19:04:42 +02:00

..

fake-profile-dir

Define four-profile production matrix with Medium default

2026-08-22 11:30:38 +02:00

benchmark_vision.py

Offload vision projector to RTX 3060

2026-08-22 17:19:56 +02:00

fake-llama-profile.sh

Define four-profile production matrix with Medium default

2026-08-22 11:30:38 +02:00

fake-systemctl.sh

Expand router into reproducible local AI platform

2026-08-20 12:56:53 +02:00

flux_benchmark_run.sh

Define four-profile production matrix with Medium default

2026-08-22 11:30:38 +02:00

flux_gpu_benchmark.py

router: Bildgenerierung mit FLUX.2 [klein] 4B Base (GPU-Hotswap)

2026-08-19 08:51:37 +02:00

flux_gpu_benchmark.sh

Define four-profile production matrix with Medium default

2026-08-22 11:30:38 +02:00

flux_offload_benchmark.py

router: Bildgenerierung mit FLUX.2 [klein] 4B Base (GPU-Hotswap)

2026-08-19 08:51:37 +02:00

flux_quality_compare.py

router: Bildgenerierung mit FLUX.2 [klein] 4B Base (GPU-Hotswap)

2026-08-19 08:51:37 +02:00

flux_steps_benchmark.py

router: Bildgenerierung mit FLUX.2 [klein] 4B Base (GPU-Hotswap)

2026-08-19 08:51:37 +02:00

mock_image_worker.py

router: Bildgenerierung mit FLUX.2 [klein] 4B Base (GPU-Hotswap)

2026-08-19 08:51:37 +02:00

mock_stt_worker.py

router: deutsche Spracherkennung mit whisper.cpp (CPU-only)

2026-08-19 13:53:22 +02:00

mock_tts_worker.py

Cleanup: letzter Kokoro-Reference entfernt

2026-08-19 22:02:59 +02:00

mock_upstream.py

Replace vision hotswap with native multimodal profiles

2026-08-20 14:24:56 +02:00

QWEN38-FINAL-ACCEPTANCE-v1.json

Add final Qwen3.8 runtime and uncensored profile

2026-08-23 00:10:53 +02:00

run_qwen38_abliterated_final.py

Add final Qwen3.8 runtime and uncensored profile

2026-08-23 00:10:53 +02:00

run_qwen38_abliterated_tuning.py

Add final Qwen3.8 runtime and uncensored profile

2026-08-23 00:10:53 +02:00

run_qwen38_final_acceptance.py

Add final Qwen3.8 runtime and uncensored profile

2026-08-23 00:10:53 +02:00

run_qwen38_pure_256k.py

Validate Pure quant for 256K Ultra profile

2026-08-22 11:01:13 +02:00

test_chunked.py

Rebuild profile router as secure recoverable v2

2026-08-20 13:41:00 +02:00

test_hass_mcp_yaml_guard.py

Add native chat background and guarded HA YAML access

2026-08-22 22:12:14 +02:00

test_local.sh

Define four-profile production matrix with Medium default

2026-08-22 11:30:38 +02:00

test_multipart.py

router: Multipart-Parser auf email.parser.BytesParser umgestellt

2026-08-19 14:37:24 +02:00

test_openwebui_filters.py

Add automatic per-request MCP selection

2026-08-23 19:04:42 +02:00

test_platform_context_mcp.py

Add self-maintaining Athena platform context MCP

2026-08-23 17:31:05 +02:00

test_profile_controller.py

Add RTX 5080 FLUX hot-swap worker

2026-08-22 17:40:32 +02:00

test_router_support.py

Add final Qwen3.8 runtime and uncensored profile

2026-08-23 00:10:53 +02:00

test_sonarr_release_grab.py

Clarify Sonarr search download behavior

2026-08-23 14:50:08 +02:00