ImageApi

Files

T

Cameron Cordes 31904fef80 Raise chat truncation default num_ctx to 32k, env-overridable

The history-truncation budget assumed an 8192-token context whenever a
chat request omitted num_ctx, while the llama-swap chat slots serve
20k-131k. Replayed transcripts past ~6k tokens were silently gutted
every turn — losing conversation history and destroying llama.cpp
KV-cache prefix reuse (full SWA re-prefill per turn).

Default is now 32768 (real conversations top out around 16k), with
AGENTIC_CHAT_DEFAULT_NUM_CTX to override per deploy, floored at
headroom + 1024. Explicit per-request num_ctx still wins.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

2026-06-09 19:14:02 -04:00

apollo_client.rs

EXIF GPS write: POST /image/exif/gps via exiftool

2026-04-28 22:25:40 +00:00

backend.rs

fix: prevent hybrid mode from leaking OpenRouter model to local llamacpp client

2026-05-26 09:55:16 -04:00

clip_client.rs

clip-search: fmt + clippy clamp + test AppState arg

2026-05-15 16:10:52 -04:00

daily_summary_job.rs

style: cargo fmt drift