From 3a970efd6278dbb95fa474921d3bb20668cac146 Mon Sep 17 00:00:00 2001
From: Joey Yakimowich-Payne <jyapayne@pm.me>
Date: Sun, 15 Mar 2026 18:55:08 -0600
Subject: [PATCH] fix: increase httpx read timeout to 600s for large context
 SSE streams

With 200k+ token contexts, Anthropic can take 60+ seconds for time
to first token. The 300s timeout was too aggressive for SSE reads
during long thinking phases.
---
 src/mnemosyne/gateway.py | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)

diff --git a/src/mnemosyne/gateway.py b/src/mnemosyne/gateway.py
index 7aa385d..a565a3c 100644
--- a/src/mnemosyne/gateway.py
+++ b/src/mnemosyne/gateway.py
@@ -1621,10 +1621,12 @@ def create_app(
     clients.update(
         {
             "anthropic": httpx.Client(
-                base_url=anthropic_upstream, timeout=httpx.Timeout(300.0, connect=30.0)
+                base_url=anthropic_upstream,
+                timeout=httpx.Timeout(600.0, connect=30.0, read=600.0),
             ),
             "openai": httpx.Client(
-                base_url=openai_upstream, timeout=httpx.Timeout(300.0, connect=30.0)
+                base_url=openai_upstream,
+                timeout=httpx.Timeout(600.0, connect=30.0, read=600.0),
             ),
         }
     )