search

janhq / cortex.cpp

Local AI API Platform
https://cortex.so
Apache License 2.0
2.09k stars 119 forks source link

planning: Cortex CLI should be a layer on top of the API Server #1492

Closed vansangpfiev closed 3 weeks ago

vansangpfiev commented 4 weeks ago

Sub-ticket from Epic: https://github.com/janhq/cortex.cpp/issues/1415

Problem Statement

Cortex CLI does not use the API Server in all commands

Feature Idea

Cortex CLI should be a layer on top of the API Server. This is quite straight forward except for cortex pull and cortex engines install. We use libcurl UI/UX for download progress for above CLI commands and get the models, engines directly from Github, HF, ... The execution via /v1/models/pull and /v1/engines/install requires a lot of changes in UI/UX and logic (will need websockets to track download progress). Related ticket: https://github.com/janhq/cortex.cpp/issues/1423

For cortex pull and cortex engines install using API server, can we move them into v1.0.2 instead of v1.0.1?

cc: @dan-homebrew @0xSage @gabrielle-ong @namchuai @nguyenhoangthuan99

vansangpfiev commented 4 weeks ago

Another thing: because CLI is a layer on top of the API server, CLI will start the Server if it is not started, before executing any command.

gabrielle-ong commented 3 weeks ago

Marking as complete, as a subtask under #1415