Release v0.2.2 · ENTERPILOT/GoModel

Changelog

8e7aa5b feat(cache): add opt-in response cache partitioning by user path (#1352)
74c9aea feat(dashboard): search every filter by subsequence (#1348)
0f9374c feat(live): show every replica's traffic in the live log view (#1341)
01be17e feat(mcp): resume downstream MCP sessions on any replica (#1340)
08b94cf feat(providers): add a billable flag so budgets and billing skip bring-your-own-key usage (#1349)
5360e8c feat(ratelimit): enforce limits across replicas sharing a database (#1339)
d3ab69f feat(virtualmodels): set request parameters per virtual model (#1353)

70567a7 fix(cache): hash every output-shaping field in the semantic cache key (#1356)
27c9ac5 fix(messages): apply slowdown to natively forwarded Anthropic requests (#1354)
888abba fix(openai): render Responses reasoning as chat reasoning_content (#1343)
c4bc4a4 fix(providers): keep a provider's routes while its credentials are replaced (#1338)
4e2af61 fix(providers): keep failed model discovery from opening the circuit breaker (#1346)
c02e432 fix(providers): narrow allowed_tools for compatible upstreams that ignore it (#1273)
d06adba fix(requestparams): skip chat format keys on the Responses API (#1380)
c6d309b fix(server): answer connections accepted just before a reload's shutdown (#1355)
ebf1711 fix(server): send response feedback for translated non-streaming messages (#1358)
443ef14 fix: close the 0.2.2 pre-release findings (#1359)

Full Changelog: v0.2.1...v0.2.2