simbarag

Files

T

ryan b361377088 Add read timeouts to llama-server clients so hangs fail fast

The OpenAI client defaults to a 600s read timeout, so a wedged
llama-server blocked requests for 10 minutes and never raised, which
meant with_fallbacks() to OpenAI never triggered. Set a per-chunk read
timeout (120s, configurable via LLAMA_TIMEOUT_SECONDS) plus one retry
on the ChatOpenAI and LLMClient clients. Context is preserved because
conversation history is rebuilt from Postgres on every request.

2026-09-11 17:05:10 -04:00

__init__.py

Persist streamed assistant answers even if the client disconnects

2026-09-10 01:33:44 -04:00

agents.py

Add read timeouts to llama-server clients so hangs fail fast

2026-09-11 17:05:10 -04:00

logic.py

Sort conversations by last activity

2026-09-10 01:33:44 -04:00

memory.py

Add user memory feature for cross-conversation recall