Feature Request: Native LM Studio Integration (like Ollama)

Feature Request: Native LM Studio Integration

Description

Is your feature request related to a problem? Please describe.

Currently, VS Code has native integration for Ollama (ollama launch vscode command, built-in model picker, official docs), but LM Studio users must rely on a third-party extension with manual configuration.

As local LLM adoption grows, having multiple local runner options is important for users. LM Studio has different models, better GPU utilization on some systems, and an OpenAI-compatible API that makes integration straightforward.

Describe the solution you would like

Add native LM Studio support similar to Ollama:

1. CLI Command: lms launch vscode (mirrors ollama launch vscode)
2. Settings UI: LM Studio option in Language Models settings
3. Model Picker: "LM Studio" provider alongside "Ollama"
4. Auto-detection: VS Code detects lms CLI and running server
5. Documentation: Official docs at code.visualstudio.com/docs/copilotchat/local-models

Describe alternatives you have considered

• Current third-party extension: NullSetIndustries.lmstudio-byok-chat-provider (892 installs)
  • Not officially supported
  • Manual configuration required
  • Not discoverable in VS Code docs
  • Unclear maintenance/longevity

Additional context

Why LM Studio?

• OpenAI-compatible API: Easy integration (same as Ollama effort)
• Growing adoption: 892+ installs for unofficial extension (organic growth)
• Different model zoo: Some models available on LM Studio but not Ollama
• Headless deployment: llmster daemon for servers/CI
• Enterprise-friendly: LM Studio has commercial support options

Technical Details

LM Studio uses OpenAI-compatible API at http://localhost:1234 by default.

Proposed integration would mirror Ollama pattern:

• Detect lms CLI command
• Start LM Studio server if not running
• Register models in Copilot Chat model picker
• Add settings UI for configuration

References

• Ollama Integration: https://docs.ollama.com/integrations/vscode
• LM Studio Extension: https://marketplace.visualstudio.com/items?itemName=NullSetIndustries.lmstudio-byok-chat-provider
• LM Studio Docs: http://lmstudio.ai/docs/developer
• LM Studio SDK: https://github.com/lmstudio-ai/lmstudio-js

Implementation Effort

Estimated: 4-6 weeks to MVP

• Phase 1: CLI detection + lms launch vscode command (2-3 weeks)
• Phase 2: Settings UI + model picker (1-2 weeks)
• Phase 3: Documentation (1 week)

Complexity: Low-Medium (LM Studio uses OpenAI-compatible API)

---

Community Interest

LM Studio users who want this: Please react with 👍 or comment below!

Use cases:

• Privacy-focused development (no cloud API calls)
• Custom fine-tuned models
• Offline development
• Cost savings (no API fees)
• Enterprise compliance (data sovereignty)

[vs-code-engineering]

This feature request is now a candidate for our backlog. The community has 60 days to upvote the issue. If it receives 20 upvotes we will move it to our backlog. If not, we will close it. To learn more about how we handle feature requests, please see our documentation.

Happy Coding!

[arty-hlr]

Instead of adding each different local provider, maybe the feature request would be for them to support openai compatible endpoints out of the box instead of having to use the OAI copilot extension on top of GitHub copilot.

[Roxxust]

Instead of adding each different local provider, maybe the feature request would be for them to support openai compatible endpoints out of the box instead of having to use the OAI copilot extension on top of GitHub copilot.

this might be more appropriate supporting either the v1/responses or v1/chat/completions, alternatively lms specifically has a REST /api/v1 endpoint which might be more beneficial for implementation if only implementing for LMS, https://lmstudio.ai/docs/developer/rest
https://lmstudio.ai/docs/developer/openai-compat

[vs-code-engineering]

🙂 This feature request received a sufficient number of community upvotes and we moved it to our backlog. To learn more about how we handle feature requests, please see our documentation.

Happy Coding!