Oct 2 20 days ago

1. 06:36

   card: v3 typed outline, whole-project ingestion, citation audit

   GhostScientist
   Verified signature

   1290884
2. 06:24

   v3.1: audit vs full tree; depth + per-section citation rules in page prompts

   GhostScientist
   Verified signature

   1bf3122
3. 05:59

   v3: raise llama-server context to 32768

   GhostScientist
   Verified signature

   2393466
4. 05:59

   v3 UI: page-type badges, citation-audit display

   GhostScientist
   Verified signature

   68a234a
5. 05:59

   v3 backend: whole-project ingestion, typed adaptive pages, citation audit

   GhostScientist
   Verified signature

   309f244
6. 05:45

   card: describe multi-page wiki flow and L4 measurements

   GhostScientist
   Verified signature

   5276486
7. 05:39

   v2 UI: sidebar navigation, per-page streaming, TOC, export

   GhostScientist
   Verified signature

   72a91d5
8. 05:39

   v2 backend: outline pass + per-page streaming wiki

   GhostScientist
   Verified signature

   179cac6
9. 05:31

   heartbeat events every 15s during prompt eval; queue-based streaming

   GhostScientist
   Verified signature

   f44d582
10. 05:17

    fix: llama-server lives in /app in the llama.cpp image, not on PATH

    GhostScientist
    Verified signature

    1746954
11. 05:12

    web UI: streaming markdown + mermaid

    GhostScientist
    Verified signature

    9917e6a
12. 05:12

    FastAPI backend: GitHub fetch/rank/prompt + NDJSON stream

    GhostScientist
    Verified signature

    e32893f
13. 05:12

    start script: llama-server + health wait + uvicorn

    GhostScientist
    Verified signature

    44c978f
14. 05:12

    Dockerfile: llama.cpp server-cuda + python venv + baked GGUF

    GhostScientist
    Verified signature

    c922f8e
15. 05:12

    Space card with docker metadata

    GhostScientist
    Verified signature

    716e517
16. 05:11

    initial commit

    GhostScientist
    Verified signature

    d21ceb5