Local models for agents now ignored in PyCharm 2026.1.1 by Cline and Goose?
Previously, I was able to use AI stacks with Cline or Goose agents, and local models from both LM Studio and Ollama, but it seems that functionality is now broken. In Cline and Goose, neither Ollama nor LM Studio is now a provider option. The provider is still set-up in Third-Party providers, but it seems to be ignored.
Basically, all three, PyCharm, and both agents, were ‘upgraded’ since April 2026, so I don't know where the fault might lie. I might try wiping and a clean install, but I figured Ids ask here first. I did not find any clear matches in the issue database .
Please sign in to leave a comment.
I have been seeing something similar after the recent updates, so it may not be an issue with your local setup alone. If Ollama and LM Studio are still running correctly and their local APIs are accessible, the problem could be related to changes in PyCharm 2026.1.1 or updates to the Cline and Goose plugins. Since both agents no longer list Ollama or LM Studio as provider options despite them being configured under Third-Party Providers, it seems possible that the latest versions are no longer reading those settings correctly or have changed how local providers are integrated. Before doing a complete reinstall, I would check the plugin changelogs and verify whether local model support has been modified, deprecated, or affected by a known bug in the latest releases.
https://intellij-support.jetbrains.com/hc/en-us/community/posts/35567162209938-Local-models-for-agents-now-ignored-in-PyCharm-2026-1-1-by-Cline-and-Geometry Dash Lite-Goose
hi bro .l can help you ,Root cause of the problem and complete step-by-step fix (adapted for PyCharm + Cline + Goose upgrade after April 2026)
1. Core reason (3 breaking changes brought by the major version update in April)
Goose 2026.04 architecture reconfiguration
On April 24th, the built-in llama.cpp local inference was updated, and the native Ollama/LM Studio independent Provider entry was removed, no longer showing separate dropdown options; old third-party configurations would be ignored by the UI, and the "OpenAI Compatible General Custom Provider" must be used to access the local service.
Cline's new classification adjustment
No longer separately listing Ollama and LM Studio, they are unified under the OpenAI Compatible (OpenAI compatible) category. Damaged old configuration cache will cause the dropdown not to display the local model list.Football Bros Unblocked
PyCharm 2026.1.x JetBrains AI plugin isolation rule change
IDE's global third-party LLM configuration is completely isolated from the Cline/Goose plugin configuration. The Ollama/LM Studio you filled in in the IDE settings will not be synchronized to the plugin, and the plugin needs to be configured separately; at the same time, after the upgrade, the plugin cache file is damaged, and the old Provider entry is invalid but will not be automatically cleaned up.