Mirrored from upstream 1jehuang/jcode — Issue #177 by @rohith-rnair
Original state: open
Created: 2026-05-08T19:31:54Z · Updated: 2026-05-12T22:55:31Z
This issue is an auto-mirrored copy. Comments and edits here are local to quangdang46/jcode — do not expect them to propagate upstream.
I really want to see JCode succeed because I think the overall direction is promising. I’ve spent several hours trying to set it up on macOS and gave it multiple chances, but the current experience has been extremely frustrating, especially around providers and models.
I consistently ran into issues where providers could not be configured properly, models were unavailable, switching models did not work, and even the default provider/models frequently failed. Sometimes the app works, but most of the time it feels unreliable and unpredictable.
There are also some concerning behaviors around permissions and telemetry. In a few cases, actions happened without clear confirmation prompts, including file/folder modifications. The amount of telemetry being collected also feels excessive and should be communicated more transparently.
I’m sharing this because I genuinely think the project has potential, but the basics need to be stable first. Provider setup, model management, reliability, permissions, and configuration UX all need significant improvement. Right now, the experience feels very unfinished, and it would help to be more transparent about what is currently stable versus experimental.
I really want to see JCode succeed because I think the overall direction is promising. I’ve spent several hours trying to set it up on macOS and gave it multiple chances, but the current experience has been extremely frustrating, especially around providers and models.
I consistently ran into issues where providers could not be configured properly, models were unavailable, switching models did not work, and even the default provider/models frequently failed. Sometimes the app works, but most of the time it feels unreliable and unpredictable.
There are also some concerning behaviors around permissions and telemetry. In a few cases, actions happened without clear confirmation prompts, including file/folder modifications. The amount of telemetry being collected also feels excessive and should be communicated more transparently.
I’m sharing this because I genuinely think the project has potential, but the basics need to be stable first. Provider setup, model management, reliability, permissions, and configuration UX all need significant improvement. Right now, the experience feels very unfinished, and it would help to be more transparent about what is currently stable versus experimental.