Inference VS Code Extension

Official VS Code extension for the Inference programming language.

Features

Syntax Highlighting

Full syntax highlighting support for Inference language constructs:

• Keywords: fn, struct, enum, type, const, let, pub, mut, spec, external
• Control Flow: if, else, loop, break, return, assert
• Non-deterministic Constructs: forall, exists, assume, unique, @ (uzumaki)
• Primitive Types: i8, i16, i32, i64, u8, u16, u32, u64, bool
• Literals: strings, numbers (decimal, hex, binary, octal), booleans
• Comments: line (//), documentation (///), and block (/* */)

Language Configuration

• Auto-closing brackets: {}, [], (), "", ''
• Comment toggling with Ctrl+/ (line) and Shift+Alt+A (block)
• Bracket matching and highlighting
• Code folding with // #region and // #endregion markers
• Smart indentation for blocks

File Association

• Automatically activates for .inf files
• Custom file icon for Inference source files

Language Server

The extension automatically starts the Inference language server (inference-lsp) for .inf files, providing rich language intelligence:

• Diagnostics - Compiler errors and warnings as you type
• Hover - Type information and documentation, including explanations of the non-deterministic operators (forall, exists, assume, unique, @)
• Go to Definition - Jump to symbol definitions (F12)
• Completions - Context-aware code completion
• Document Symbols - Outline view and breadcrumb navigation
• Inlay Hints - Inline type annotations

The server binary is resolved using the following priority (mirroring infs detection):

1. Custom path from inference.lsp.path setting (if set but not executable, the server is not started - no fallback)
2. Managed installation in INFERENCE_HOME/bin/inference-lsp (respects INFERENCE_HOME environment variable)
3. System PATH

If the binary is not found, the extension stays quiet: a line is logged to the "Inference" output channel and the server simply stays off until the toolchain is installed or updated. Server traces are written to the dedicated "Inference Language Server" output channel. Use Inference: Restart Language Server to pick up a new binary after an update or a settings change made outside VS Code.

Toolchain Management

The extension provides comprehensive toolchain management through integration with the infs CLI. All operations are fully automated and require no manual configuration.

Automatic Detection

On activation, the extension automatically detects your toolchain using the following priority:

1. Custom path from inference.path setting
2. Managed installation in INFERENCE_HOME/bin/infs (respects INFERENCE_HOME environment variable)
3. System PATH

The detection result is displayed in the Configuration sidebar and logged to the Output channel.

Configuration Sidebar

A dedicated Inference icon appears in the VS Code activity bar. Click it to open the Configuration view with real-time toolchain information:

Toolchain Group:

• Binary path and detection source (settings/managed/path)
• Installed version number
• INFERENCE_HOME directory location (default or custom)
• Detected platform (e.g., linux-x64, macos-arm64, windows-x64)
• Health status with diagnostic results

Settings Group:

• inference.path - Custom binary path (click to configure)
• inference.autoInstall - Auto-install prompt behavior
• inference.checkForUpdates - Automatic update checking

Interactive Actions:

• Click any path item to copy its value to clipboard
• Right-click path items to reveal in file explorer
• Click status to run doctor diagnostics
• Use the refresh button in the title bar to reload

The view automatically refreshes when settings change or after install/update operations.

Terminal Integration

The extension automatically prepends INFERENCE_HOME/bin to PATH for all VS Code integrated terminals using the EnvironmentVariableCollection API:

• New terminals immediately have infs and infc available
• Existing terminals show a relaunch indicator when the toolchain changes
• No VS Code restart required after installation or updates
• Works across all supported platforms

Status Bar

The bottom-left status bar shows real-time toolchain health. Click the status bar item to run full diagnostics via infs doctor.

Available Commands

Open Command Palette (Ctrl+Shift+P / Cmd+Shift+P):

• Inference: Install Toolchain - Download and install the latest infs release for your platform
• Inference: Update Toolchain - Check for updates and install the latest version
• Inference: Select Toolchain Version - Browse and switch between available versions
• Inference: Run Doctor - Execute comprehensive health diagnostics
• Inference: Restart Language Server - Stop and restart inference-lsp, re-resolving the binary location
• Inference: Refresh Configuration - Reload the Configuration sidebar view
• Inference: Show Output - Open the Inference output log channel
• Inference: Reset PATH Fallback Preference - Clear saved PATH fallback acceptance

A guided setup walkthrough is available via Get Started: Open Walkthrough... > Get Started with Inference.

Installation

From VS Code Marketplace

1. Open VS Code
2. Press Ctrl+P to open Quick Open
3. Type ext install inference-lang.inference
4. Press Enter

From VSIX

1. Download the .vsix file from Releases
2. In VS Code, press Ctrl+Shift+P
3. Type "Install from VSIX" and select the command
4. Choose the downloaded .vsix file

Configuration

Settings

• inference.path (string, default: "") - Custom path to the infs binary. Leave empty for automatic detection. Scope: machine (not synced across devices).
• inference.autoInstall (boolean, default: true) - Prompt to install toolchain if not found on activation.
• inference.checkForUpdates (boolean, default: true) - Automatically check for toolchain updates on activation.
• inference.lsp.enabled (boolean, default: true) - Start the Inference language server automatically. Disable to turn off all language intelligence features.
• inference.lsp.path (string, default: "") - Custom path to the inference-lsp binary. Leave empty for automatic detection. Scope: machine (not synced across devices).

Environment Variables

• INFERENCE_HOME - Override default toolchain directory (default: ~/.inference on Linux/macOS, %APPDATA%\inference on Windows — the same locations infs uses)
• INFS_DIST_SERVER - Override distribution server URL (for development/testing)

Supported Platforms

Automatic toolchain installation is supported on:

• Linux: x86_64 (glibc)
• macOS: ARM64 (Apple Silicon)
• Windows: x86_64

Other platforms can use the extension for syntax highlighting but must install the toolchain manually.

Example

/// Computes factorial using non-deterministic verification
pub fn factorial(n: i32) -> i32 {
    let mut result: i32 = 1;
    let mut i: i32 = 1;

    loop {
        if i > n {
            break;
        }
        result = result * i;
        i = i + 1;
    }

    // Verify the result using forall block
    forall {
        const witness: i32 = @;
        assume {
            const valid: bool = witness >= 0;
        }
    }

    return result;
}

What is Inference?

Inference is a programming language designed for mission-critical applications development. It includes first-class support for formal verification via translation to Rocq (Coq) and targets WebAssembly as its primary runtime platform.

Key features:

• Formal Verification: Built-in support for proofs and specifications
• Non-deterministic Programming: forall, exists, assume, unique constructs
• WebAssembly Target: Compiles to efficient WASM
• Rocq Translation: Generate Coq proofs from your code

Learn more:

• Inference Repository
• Language Specification
• Inference Book

Troubleshooting

Toolchain not detected

1. Check the Output panel (View > Output > Select "Inference")
2. Run Inference: Run Doctor to see detailed diagnostics
3. Verify inference.path setting if using a custom location
4. Try Inference: Install Toolchain to install automatically

Language server not running

1. Check the Output panel (View > Output > Select "Inference") for a "Language server" log line explaining why it did not start
2. Ensure inference.lsp.enabled is true
3. Install or update the toolchain (Inference: Install Toolchain / Inference: Update Toolchain) so inference-lsp is present in INFERENCE_HOME/bin
4. Alternatively, set inference.lsp.path to the binary location (note: if this setting points to a non-executable path, the server is not started and no fallback occurs)
5. Run Inference: Restart Language Server after fixing the binary location

Terminal commands not found

1. Close all open terminals and open a new one (Terminal > New Terminal)
2. The extension automatically adds INFERENCE_HOME/bin to PATH
3. For external terminals, add the path to your shell profile manually

Privacy

This extension does not collect telemetry, usage data, or any personal information. All toolchain operations communicate only with github.com/Inferara/inference/releases and inference-lang.org/releases.json.

Contributing

Contributions are welcome! Please see the main repository for contribution guidelines.

License

GPL-3.0 - See LICENSE for details.