Native inline suggestions
VS Code ghost text with configurable automatic and manual triggers, multiline suggestions and cancellation.
Natural inline suggestions in VS Code, powered by the Ollama model you run. No LineLoom account, no built-in analytics, and no mandatory cloud service.
Available from Open VSX Registry for compatible editors. Stable v1.0.0 release. Requires VS Code 1.96+ and your independently installed Ollama model.
Prefer manual installation? Download the LineLoom-1.0.0.vsix GitHub package and install it inside VS Code.
Focused tools for the code you're already writing—not a service you have to surrender your workflow to.
VS Code ghost text with configurable automatic and manual triggers, multiline suggestions and cancellation.
Choose your installed default model, or assign a different installed model to each supported language.
Trusted workspaces, path exclusions and explicit remote endpoint approval help limit unintended disclosure.
Reuse a short-lived suggestion when the typed prefix still matches, without asking your model again.
Control completion, models, connection checks, language settings and diagnostics from the status bar.
MIT-licensed source, documented architecture, release artifacts and automated checks in public GitHub.
Nine registered language IDs. Go, Java, C and C++ remain experimental in this release.
No LineLoom sign-up or preconfigured cloud key required. You bring the model. Windows VSIX Installer showing an error? Read the VS Code install guide ↗
Get Ollama for your operating system and start its service.
Get OllamaDownload an example model yourself. Model terms and hardware needs vary.
ollama pull qwen2.5-coder:1.5bInstall from Visual Studio Marketplace or Open VSX Registry in a compatible editor. A direct GitHub VSIX is also available.
Compare installation optionsOpen a trusted project and use “LineLoom: Quick Actions” to choose a model and check the connection.
TroubleshootingThe default endpoint is http://127.0.0.1:11434. You can configure a remote HTTPS endpoint only with explicit approval; in that case code snippets can leave your device.
LineLoom submits bounded context from the active document to your configured Ollama-compatible endpoint. On loopback, requests stay on your computer's networking stack. If you approve a remote HTTPS endpoint, the snippets travel to that server and are governed by its policies.
LineLoom has no built-in telemetry or automatic repository indexing. It excludes sensitive paths, but it does not provide comprehensive content-level secret detection. Review your settings before using confidential code.
Read the full privacy policyYes. The extension source is released under the MIT License. Ollama and individual downloaded models have separate licenses and hardware requirements.
After independently installing Ollama, a compatible model, and the extension, loopback inference can work without internet access. Fully offline setup and behavior across platforms have not yet been manually validated for stable release.
v1.0.0 is a GitHub stable release. Automated testing and a clean VS Code CLI install are verified. Actual model quality and hands-on acceptance across platforms are not independently certified.
Yes. LineLoom v1.0.0 is public on Visual Studio Marketplace and Open VSX Registry, both under publisher SanskarCodes. Open VSX currently shows no publisher verification badge; the extension is nonetheless publicly downloadable. The GitHub stable VSIX is available as another option.
Cross-file retrieval/indexing is not implemented in this release. LineLoom currently uses bounded context from the active trusted-workspace document.
Install the stable release, send feedback and help shape what's next.