Responses API history duplication and compression bugs
NousResearch/hermes-agent · cluster 56895 · agent
Data as of
Nothing open on GitHub. Fix merged in #69306. Check it covers your case.
The short answer
- Known?
- Yes. 12 related items grouped here, 3 issues already closed.
- Fix event
- A fix was merged to main in PR #69306. Check it covers your case. No release data is tracked.
- Tracker
- The maintainers' tracker marks this
partially resolved, deciding PR #69306. That is separate from the fix event. - Still open
- Nothing open on GitHub.
These issues and fixes center on the /v1/responses endpoint experiencing conversation history duplication and repeated compression loops when context exceeds model limits. Multiple related problems are addressed: compressed transcripts not being persisted correctly, history growing exponentially, and post-compression state not being properly stored or advanced.
Open issues0
Open PRs0
Fixes on record
-
#69306 fix(api): persist compressed transcripts in ResponseStore to stop re-compression loops
files:
contributors/emails/[redacted:email],contributors/emails/[redacted:email],gateway/platforms/api_server.py,tests/gateway/test_api_server.py -
#116326 Chained /v1/responses turns no longer replay earlier tool calls or double the stored history after history repair/compaction (#89891, rebuild of #70695)
files:
gateway/platforms/api_server_openai_routes.py,gateway/platforms/api_server_turn_boundary.py,tests/gateway/test_api_server_responses_history_dedupe.py,tests/gateway/test_api_server_turn_boundary.py,website/docs/user-guide/features/api-server.md
7 PRs closed without merging
From the close comments: done elsewhere 3 · partial fix 2 · duplicate 1 · merged via another change 1
-
#56933 fix(api): advance responses chain after compression · done elsewhere →
d12d8b3 -
#56996 fix(api): persist compressed responses history · duplicate →
#56933,#56895 -
#58133 fix(gateway): Responses API stores bloated context after compression · merged via another change →
1c21e96,#69306 -
#66148 fix(gateway): prune stored /v1/responses history images and stop post-compression history duplication · partial fix →
1c21e96,c7502a5,#56895,#56933,#56996,#69306,#66148,#1,#15803,#14532,#6,#9,#374151,#4338 -
#68282 fix(api): keep Responses history system-free 🩷 · partial fix →
d1ea838,#68257,#68282,#70695,#75289,#1,#15803,#14532,#6,#9,#374151,#4338
Closed issues · 3
- Closed#56895`/v1/responses` enters repeated compression loop after context exceeds model limit
- Closed#58081[Bug] Responses API: compressed transcript not stored in ResponseStore, causing re-compression loop
- Closed#68257Exponential conversation history doubling in Responses API (_build_response_conversation_history)
How this cluster was grouped
-
Link evidence (
partly structural links): some members are connected by structural evidence (shared changed files, or a closing/superseded reference); the rest are joined by an automated relevance judgement over their text. - Members outside that structural core may belong to a different underlying problem.
-
Pairwise link density
0.16: the share of item pairs that carry a link, counting a structural link in full and a relevance-only link at 0.36.1.00means every pair is linked.
Pick up a Lab task
Check whether a fix covers the open issues here, with your own agent. Bring back evidence for maintainers.
Open the Lab