0.5.0-46

Changes

• feat: expose server instance @marknguyen1302 (#1025)
• fix: engine init is broken @louis-jan (#1022)
• Resolve comment in brew PR @hiento09 (#1018)
• feat: support to change config path @marknguyen1302 (#1009)
• Feature dockerfile for cortex @hiento09 (#1014)
• chore: fallback noavx binary @louis-jan (#1013)
• fix: add used vram into system event body @louis-jan (#1012)
• chore: system resource events should include gpu system information @louis-jan (#1011)
• fix: remove uploads_folder_path @vansangpfiev (#996)
• feat: unload engine when init engine @marknguyen1302 (#1003)
• fix: binary file lookup @louis-jan (#1001)
• feat: add pagination for thread list api @marknguyen1302 (#1000)
• fix: set namespace for js module @louis-jan (#999)
• fix: increase CI timeout @louis-jan (#998)
• chore: support more configurations via direct js import @louis-jan (#997)
• Fix brew build from source error @hiento09 (#995)
• fix: hide gpu check windows @louis-jan (#993)
• Chore bump cortex-cpp to 0.5.0-34 @hiento09 (#992)
• Fix cortex-cpp rebuild @hiento09 (#994)
• fix: fix checksum undefined @marknguyen1302 (#991)
• feat: support unload engine @vansangpfiev (#989)
• cortex js allow tag x.y.z-t @hiento09 (#990)
• add build-deps for cortex-cpp binding @hiento09 (#988)
• fix: env does not work after bundled in an electron app @louis-jan (#987)
• feat: download model checksum @marknguyen1302 (#986)
• Windows uninstaller should stop cortex and remove cortex home folder for all users @hiento09 (#984)
• chore: name cortex processes @louis-jan (#983)
• fix: correct swagger port @marknguyen1302 (#980)
• fix: missing utils reference @louis-jan (#978)
• fix: force change localhost @marknguyen1302 (#977)
• chore: bump cpp version to 0.5.0-27 @marknguyen1302 (#976)
• chore: return error on pulling an existing model @louis-jan (#975)
• feat: add config cpp port @marknguyen1302 (#973)
• chore: clean logs periodically @louis-jan (#972)
• Fix codesign macos for application and installer @hiento09 (#974)
• fix: persists engine version on init @louis-jan (#969)
• feat: support multi process in cortex @marknguyen1302 (#965)
• Update the Base URL for the API reference to reflect the new URL @irfanpena (#971)
• fix: disable devtool @marknguyen1302 (#967)
• Fix codesign windows @hiento09 (#964)
• fix: create message request with string content instead of array @louis-jan (#963)
• fix: cortex engines init api with empty body should init default options @louis-jan (#962)
• fix: change default host cortex @marknguyen1302 (#961)
• feat: watch models and engines update for proper data retrieval @louis-jan (#960)
• fix: fix issue when re-download aborted model @marknguyen1302 (#959)
• Add launchpad uninstaller prerm and postrm @hiento09 (#957)
• feat: support martian nvidia engine @marknguyen1302 (#956)
• chore: delete local file when abort download @marknguyen1302 (#954)
• fix: terminate system should kill cortex process @louis-jan (#955)
• fix: start and run models are not outputting last error logs @louis-jan (#951)
• fix: remove hardcode stream @marknguyen1302 (#952)
• chore: persist context length and ngl from gguf file @louis-jan (#947)
• chore: show error from remote engine, lint @marknguyen1302 (#949)
• Fix openai api pipeline @hiento09 (#948)
• chore: destroy dangling processes on uninstall @louis-jan (#945)
• feat: support openrouter, cohere engine @marknguyen1302 (#946)
• feature: support local model pull @louis-jan (#944)
• chore: handle failed download @marknguyen1302 (#943)
• chore: specify engine version to pull @louis-jan (#942)
• bump cortex llamacpp for lla‌ma 3.1 @Van-QA (#941)

Contributor

@Van-QA, @hiento09, @irfanpena, @louis-jan, @marknguyen1302 and @vansangpfiev