• Joined on 2024-12-20
doru pushed to master at doru/intelligence-router 2026-06-18 03:12:06 +03:00
b3ac21b2c0 fix: first request no longer blocks on model switch — uses background task + SSE
doru pushed to master at doru/intelligence-router 2026-06-18 02:51:02 +03:00
45dd793b69 fix: sidecar process kill was not awaiting wait() — old server held GPU VRAM
doru pushed to master at doru/intelligence-router 2026-06-17 01:09:17 +03:00
7e9b3f43e1 fix: circuit breaker deadlock — always query sidecar for status
doru pushed to master at doru/intelligence-router 2026-06-17 00:46:07 +03:00
bcf45129f1 fix: add --host 0.0.0.0 to llama-server command
doru pushed to master at doru/intelligence-router 2026-06-17 00:32:38 +03:00
75248741e7 fix: log exceptions on primary proxy target
doru pushed to master at doru/intelligence-router 2026-06-17 00:25:42 +03:00
5c1753dfef fix: log sidecar switch failures + fix scoping bug in proxy handler
doru pushed to master at doru/intelligence-router 2026-06-17 00:16:47 +03:00
f2e62f60e6 fix: /api/show GET support, /v1 root handler, and proxy debug logging
doru pushed to master at doru/intelligence-router 2026-06-17 00:04:42 +03:00
d935339280 fix: report actual profile context size in /api/show probe endpoint
doru pushed to master at doru/intelligence-router 2026-06-16 23:54:32 +03:00
4ee85972ec fix: convert underscores to hyphens in llama-server flag names, fix n_ctx→ctx-size rename
doru pushed to master at doru/intelligence-router 2026-06-16 23:36:19 +03:00
1551c281c2 fix: move llama-server stderr log from /tmp to working dir (ReadWritePaths compat)
doru pushed to master at doru/intelligence-router 2026-06-16 03:06:49 +03:00
37fee5341e fix: capture llama-server stderr, fix YAML boolean flag conversion, reduce polling timeout
doru pushed to master at doru/intelligence-router 2026-06-16 00:10:37 +03:00
903f06c634 feat: add sync_models.py script to auto-update Hermes custom_providers from router model list
doru pushed to master at doru/intelligence-router 2026-06-15 19:38:18 +03:00
95c87a764b fix: remove non-existent models from manifest (qwen-3-8b, llama-4-maverick), add 3 newly discovered models
doru pushed to master at doru/intelligence-router 2026-06-15 19:25:59 +03:00
36abbf573e fix: unbuffer sidecar stdout so logs appear in journalctl
doru pushed to master at doru/intelligence-router 2026-06-15 19:02:11 +03:00
1e9305395e Fixed llama-server path
doru pushed to master at doru/intelligence-router 2026-06-15 18:31:33 +03:00
7e86a30bd8 fix: resolve port conflict between sidecar and llama-server
doru pushed to master at doru/intelligence-router 2026-06-15 18:22:16 +03:00
2c23faa4a1 fix: add probe endpoints and no-model fallback for Hermes Desktop compatibility
doru pushed to master at doru/intelligence-router 2026-06-15 18:10:55 +03:00
af12370632 changed llama-server location
doru pushed to master at doru/intelligence-router 2026-06-15 16:23:09 +03:00
1ef8a497f6 fix: update docker-compose.yml SIDECAR_URL to port 8080
doru pushed to master at doru/intelligence-router 2026-06-15 16:19:26 +03:00
45417068ae fix: change sidecar port from 8081 to 8080