# Support local LLM's **URL:** \ **Category:** Feature Requests\ **Created:** [October 16, 2023, 7:17pm UTC](https://forum.cursor.com/t/support-local-llms/1099 "2023-10-16T19:17:15Z")\ **Posts on this page:** 13\ **Page:** 3
**Author:** ![Taidan](https://avatars.discourse-cdn.com/v4/letter/t/5f8ce5/32.png) [@Taidan](https://forum.cursor.com/u/Taidan)\ **Post date:** [June 25, 2024, 10:45pm UTC](https://forum.cursor.com/t/support-local-llms/1099/42 "2024-06-25T22:45:36Z")
Yes. Please allow us to use local models, even if it’s a paid feature ---
**Author:** ![dandv](https://sea3.discourse-cdn.com/cursor1/user_avatar/forum.cursor.com/dandv/32/2018_2.png) [@dandv](https://forum.cursor.com/u/dandv)\ **Post date:** [June 28, 2024, 2:14am UTC](https://forum.cursor.com/t/support-local-llms/1099/43 "2024-06-28T02:14:17Z")
> [@tschwarz](#): > > No, [Continue is] noticeably worse In the meantime Continue published a post about [using Mistral AI’s Codestral 22B, DeepSeek Coder 6.7B, and Llama 3 8B](https://blog.continue.dev/ollama-code-assistant/). ---
**Author:** ![tschwarz](https://sea3.discourse-cdn.com/cursor1/user_avatar/forum.cursor.com/tschwarz/32/19807_2.png) [@tschwarz](https://forum.cursor.com/u/tschwarz)\ **Post date:** [June 28, 2024, 2:58am UTC](https://forum.cursor.com/t/support-local-llms/1099/44 "2024-06-28T02:58:58Z")
if i buy a new laptop i’d definitely be looking for something with enough ram to run some of these locally, pretty incredible! ---
**Author:** ![ark-m](https://sea3.discourse-cdn.com/cursor1/user_avatar/forum.cursor.com/ark-m/32/2677_2.png) [@ark-m](https://forum.cursor.com/u/ark-m)\ **Post date:** [July 1, 2024, 11:37pm UTC](https://forum.cursor.com/t/support-local-llms/1099/45 "2024-07-01T23:37:13Z")
Managed to connect a local Coder-V2 by using ollama and cloudflared tunnel. 16b model is super slow on my Mac M1 Pro, takes a minute to process any request, even the simplest ones. But the main issue it breaks the code, injects some weird markdowns, cuts lines, and so on. Seems like it needs some additional instructions and prompts tuning on the Cursor side. So guys, technically it’s possible but as of now it’s far from ready-to-use state. ---
**Author:** ![WEBA11Y](https://avatars.discourse-cdn.com/v4/letter/w/4da419/32.png) [@WEBA11Y](https://forum.cursor.com/u/WEBA11Y)\ **Post date:** [July 15, 2024, 5:44pm UTC](https://forum.cursor.com/t/support-local-llms/1099/46 "2024-07-15T17:44:33Z")
I am using LM Studio. I get this base url [http://localhost:1234/v1](http://localhost:1234/v1) and I used it on cursor but not working. Here is the curl ```auto curl http://localhost:1234/v1/chat/completions -H "Content-Type: application/json" -H "Authorization: Bearer lm-studio" -d "{ \"messages\": [{ \"role\": \"system\", \"content\": \"You are a test assistant.\" }, { \"role\": \"user\", \"content\": \"Testing. Just say hi and nothing else.\" }], \"model\": \"bartowski/DeepSeek-Coder-V2-Lite-Instruct-GGUF\" }" ``` Please someone tell me how to use local LLM on Cursor. ---
**Author:** ![deanrie](https://sea3.discourse-cdn.com/cursor1/user_avatar/forum.cursor.com/deanrie/32/1263_2.png) [@deanrie](https://forum.cursor.com/u/deanrie)\ **Post date:** [July 16, 2024, 8:01am UTC](https://forum.cursor.com/t/support-local-llms/1099/47 "2024-07-16T08:01:34Z")
You must have an external address. You can configure this through various proxying services or find something ready-made. Like this one works, but I’m not sure [GitHub - kcolemangt/llm-router: Access models from OpenAI, Groq, local Ollama, and others by setting llm-router as Cursor's Base URL](https://github.com/kcolemangt/llm-router) ---
**Author:** ![labcode](https://avatars.discourse-cdn.com/v4/letter/l/3bc359/32.png) [@labcode](https://forum.cursor.com/u/labcode)\ **Post date:** [August 10, 2024, 4:46am UTC](https://forum.cursor.com/t/support-local-llms/1099/48 "2024-08-10T04:46:07Z")
This needs to happen. Let me experiment using Cursor with Llama 3.1 I have on my own machine. ---
**Author:** ![briprowe](https://sea3.discourse-cdn.com/cursor1/user_avatar/forum.cursor.com/briprowe/32/21269_2.png) [@briprowe](https://forum.cursor.com/u/briprowe)\ **Post date:** [January 21, 2025, 10:03pm UTC](https://forum.cursor.com/t/support-local-llms/1099/50 "2025-01-21T22:03:28Z")
I won’t be using cursor for serious work until this is in place. ---
**Author:** ![dibun](https://avatars.discourse-cdn.com/v4/letter/d/e79b87/32.png) [@dibun](https://forum.cursor.com/u/dibun)\ **Post date:** [January 29, 2025, 10:23pm UTC](https://forum.cursor.com/t/support-local-llms/1099/51 "2025-01-29T22:23:50Z")
Are there any plans to support [Model Catalog - LM Studio](https://lmstudio.ai/models) ---
**Author:** ![dandv](https://sea3.discourse-cdn.com/cursor1/user_avatar/forum.cursor.com/dandv/32/2018_2.png) [@dandv](https://forum.cursor.com/u/dandv)\ **Post date:** [March 1, 2025, 5:34am UTC](https://forum.cursor.com/t/support-local-llms/1099/52 "2025-03-01T05:34:58Z")
@Jakob since this reply, DeepSeek-R1 was released, so the argument that “the state of OSS at the moment” would lead to a “much worse product experience”, no longer holds. Can this topic be marked as unanswered? ---
**Author:** ![adamcodes716](https://sea3.discourse-cdn.com/cursor1/user_avatar/forum.cursor.com/adamcodes716/32/61407_2.png) [@adamcodes716](https://forum.cursor.com/u/adamcodes716)\ **Post date:** [June 5, 2025, 9:32pm UTC](https://forum.cursor.com/t/support-local-llms/1099/53 "2025-06-05T21:32:17Z")
How did you manage to do this? I am using a cloudflare tunnel to access my ollama service. A curl command accesses it with no problem, but I can’t get commands from cursor to show up in the ollama console. ---
**Author:** ![user1007](https://avatars.discourse-cdn.com/v4/letter/u/a6a055/32.png) [@user1007](https://forum.cursor.com/u/user1007)\ **Post date:** [January 11, 2026, 7:53am UTC](https://forum.cursor.com/t/support-local-llms/1099/54 "2026-01-11T07:53:25Z")
What’s the status here? I’m finding that right as I get jamming on my projects I’m getting shafted with the usage limit. Meaning, 1. Auto is too expensive and/or 2. Running your own local LLMs needs to be an option. ---
**Author:** ![dandv](https://sea3.discourse-cdn.com/cursor1/user_avatar/forum.cursor.com/dandv/32/2018_2.png) [@dandv](https://forum.cursor.com/u/dandv)\ **Post date:** [June 15, 2026, 6:42am UTC](https://forum.cursor.com/t/support-local-llms/1099/55 "2026-06-15T06:42:08Z")
> [@user1007](#): > > 1. Auto is too expensive and/or 2. Running your own local LLMs needs to be an option. Take a look at [headroom](https://github.com/chopratejas/headroom) for reducing input token count. > [@user1007](#): > > 2. Running your own local LLMs needs to be an option. I don’t think Cursor’s business model warrants them moving in that direction. For GUI-based agentic development using local LLMs, take a look at Zed or OpenCode, if you don’t need [integrated browser support](https://cursor.com/docs/agent/tools/browser) via CDP, and you’re fine instructing your LLM explicitly to instrument code for [debugging](https://cursor.com/docs/agent/debug-mode). [Previous page](https://forum.cursor.com/t/support-local-llms/1099.md?page=2)