diff --git a/.gitignore b/.gitignore
index 59cbdc3..5de1d86 100644
--- a/.gitignore
+++ b/.gitignore
@@ -1,5 +1,6 @@
 
 zig-cache/
+.zig-cache/
 zig-out/
 /release/
 /debug/
diff --git a/README.md b/README.md
index 0ccd0ae..97a5853 100644
--- a/README.md
+++ b/README.md
@@ -58,28 +58,12 @@ fantastic projects from the Zig community:
 
 ## Installing
 
-Requires [zig v0.12.x](https://ziglang.org).
-1.  Initialize your project repository:
-    ```bash
-    git init
-    ```
-2.  Create a `libs` directory inside the root of your project:
-    ```bash
-    mkdir libs
-    ```
-3.  Add this library as a submodule of your project:
-    ```bash
-    git submodule add https://github.com/TSxo/zli libs/zli
-    ```
-4.  Inside your `build.zig`, bring in the `zli` module and add it as an import
+Requires [zig v0.14.x](https://ziglang.org).
+1.  `zig fetch --save "git+https://github.com/TSxo/zli#v0.1.1"`
+2.  Inside your `build.zig`, bring in the `zli` module and add it as an import
     to your exe:
     ```zig
-    const module = b.addModule("zli", .{
-        .root_source_file = .{
-            .path = "libs/zli/src/zli.zig",
-        },
-    });
-
+    // Your typical exe
     const exe = b.addExecutable(.{
         .name = "project",
         .root_source_file = b.path("src/main.zig"),
@@ -87,9 +71,14 @@ Requires [zig v0.12.x](https://ziglang.org).
         .optimize = optimize,
     });
 
-    exe.root_module.addImport("zli", module);
+    // Now, add zli to the exe
+    const zli = b.dependency("zli", .{
+        .target = target,
+        .optimize = optimize,
+    });
+    exe.root_module.addImport("zli", zli.module("zli"));
     ```
-5.  `zli` is now usable in your project:
+3.  `zli` is now usable in your project:
     ```zig
     const zli = @import("zli");
     ```
diff --git a/build.zig b/build.zig
index 0419eb4..e9c5787 100644
--- a/build.zig
+++ b/build.zig
@@ -5,13 +5,11 @@ pub fn build(b: *std.Build) void {
     const optimize = b.standardOptimizeOption(.{});
 
     const module = b.addModule("zli", .{
-        .root_source_file = .{
-            .path = "src/zli.zig",
-        },
+        .root_source_file = b.path("src/zli.zig"),
     });
 
     const main_tests = b.addTest(.{
-        .root_source_file = .{ .path = "src/test.zig" },
+        .root_source_file = b.path("src/test.zig"),
         .optimize = optimize,
         .target = target,
     });
@@ -24,9 +22,7 @@ pub fn build(b: *std.Build) void {
     inline for (.{ "subcommands", "args", "simple" }) |name| {
         const example = b.addExecutable(.{
             .name = name,
-            .root_source_file = .{
-                .path = b.fmt("example/{s}.zig", .{name}),
-            },
+            .root_source_file = b.path(b.fmt("example/{s}.zig", .{name})),
             .target = target,
             .optimize = optimize,
         });
diff --git a/build.zig.zon b/build.zig.zon
new file mode 100644
index 0000000..cb720e3
--- /dev/null
+++ b/build.zig.zon
@@ -0,0 +1,44 @@
+.{
+    .name = .zli,
+    .version = "0.1.1",
+    .fingerprint = 0x5b01c9c38221e7a5, // Changing this has security and trust implications.
+    .minimum_zig_version = "0.14.0",
+    .dependencies = .{
+        // See `zig fetch --save <url>` for a command-line interface for adding dependencies.
+        //.example = .{
+        //    // When updating this field to a new URL, be sure to delete the corresponding
+        //    // `hash`, otherwise you are communicating that you expect to find the old hash at
+        //    // the new URL. If the contents of a URL change this will result in a hash mismatch
+        //    // which will prevent zig from using it.
+        //    .url = "https://example.com/foo.tar.gz",
+        //
+        //    // This is computed from the file contents of the directory of files that is
+        //    // obtained after fetching `url` and applying the inclusion rules given by
+        //    // `paths`.
+        //    //
+        //    // This field is the source of truth; packages do not come from a `url`; they
+        //    // come from a `hash`. `url` is just one of many possible mirrors for how to
+        //    // obtain a package matching this `hash`.
+        //    //
+        //    // Uses the [multihash](https://multiformats.io/multihash/) format.
+        //    .hash = "...",
+        //
+        //    // When this is provided, the package is found in a directory relative to the
+        //    // build root. In this case the package's hash is irrelevant and therefore not
+        //    // computed. This field and `url` are mutually exclusive.
+        //    .path = "foo",
+        //
+        //    // When this is set to `true`, a package is declared to be lazily
+        //    // fetched. This makes the dependency only get fetched if it is
+        //    // actually used.
+        //    .lazy = false,
+        //},
+    },
+
+    .paths = .{
+        "build.zig",
+        "build.zig.zon",
+        "src",
+        "LICENSE",
+    },
+}
diff --git a/example/simple.zig b/example/simple.zig
index 20232d8..b19f621 100644
--- a/example/simple.zig
+++ b/example/simple.zig
@@ -46,7 +46,10 @@ pub fn main() !void {
             verbose: bool = false,
             positional: struct {
                 p1: []const u8,
-                p2: []const u8,
+                // optional, with default value
+                p2: []const u8 = "hello",
+                // optional, with default null
+                p3: ?[]const u8 = null,
             },
 
             pub const aliases = .{
@@ -95,6 +98,7 @@ pub fn main() !void {
             try writer.print("\tVerbose: {}\n", .{values.verbose});
             try writer.print("\tPos 1: {s}\n", .{values.positional.p1});
             try writer.print("\tPos 2: {s}\n", .{values.positional.p2});
+            try writer.print("\tPos 3: {?s}\n", .{values.positional.p3});
         },
         .required => |values| {
             try writer.print("Required:\n", .{});
diff --git a/src/arg.zig b/src/arg.zig
index bf71cb7..ca0d2ba 100644
--- a/src/arg.zig
+++ b/src/arg.zig
@@ -9,12 +9,12 @@ const strings = @import("./strings.zig");
 /// Asserts the type of a given value is valid.
 pub fn assert_valid_value_type(comptime T: type) void {
     comptime {
-        if (T == []const u8 or T == [:0]const u8 or @typeInfo(T) == .Int) {
+        if (T == []const u8 or T == [:0]const u8 or @typeInfo(T) == .int) {
             return;
         }
 
-        if (@typeInfo(T) == .Enum) {
-            const info = @typeInfo(T).Enum;
+        if (@typeInfo(T) == .@"enum") {
+            const info = @typeInfo(T).@"enum";
             assert(info.is_exhaustive);
             assert(info.fields.len >= 2);
             return;
@@ -33,14 +33,13 @@ pub fn assert_valid_value_type(comptime T: type) void {
 /// assert(std.mem.eql(name("-v"), "v"));
 /// ```
 pub fn name(arg: []const u8) []const u8 {
-    assert(arg.len > 1);
-    assert(arg[0] == '-');
+    if (arg.len <= 1) return "";
+    if (arg[0] != '-') return "";
 
     var from: usize = 1;
     var to: usize = arg.len;
 
-    if (arg[1] == '-') {
-        assert(arg.len > 2);
+    if (arg.len > 2 and arg[1] == '-') {
         from = 2;
     }
 
@@ -66,9 +65,12 @@ test name {
 /// assert(std.mem.eql(unparsed_value("-n=5"), "5"));
 /// ```
 fn unparsed_value(arg: []const u8) []const u8 {
-    const split_at = strings.index_of(arg, "=").? + 1;
-    assert(arg.len > split_at);
-    return arg[split_at..];
+    if (strings.index_of(arg, "=")) |pos_of_equ| {
+        const split_at = pos_of_equ + 1;
+        if (arg.len > split_at)
+            return arg[split_at..];
+    }
+    return "";
 }
 
 test unparsed_value {
@@ -86,7 +88,9 @@ test unparsed_value {
 pub fn parse_arg(comptime T: type, arg: []const u8) T {
     const arg_name = name(arg);
     const val = unparsed_value(arg);
-    assert(val.len > 0);
+    if (val.len == 0) {
+        fatal("Could not parse argument `{s}`: value length is 0. Did you forget the `=`? (like: `-{s}=`)", .{ arg_name, arg_name });
+    }
     return parse_value(T, arg_name, val);
 }
 
@@ -99,16 +103,21 @@ pub fn parse_arg(comptime T: type, arg: []const u8) T {
 /// ```
 pub fn parse_value(comptime T: type, arg_name: []const u8, arg_value: []const u8) T {
     if (T == bool) return true;
-    assert(arg_value.len > 0);
+
+    // this error is usually handled in parse_arg() already but checking for it
+    // here doesn't harm.
+    if (arg_value.len == 0) {
+        fatal("Value for argument `{s}` has zero length!", .{arg_name});
+    }
 
     const V = switch (@typeInfo(T)) {
-        .Optional => |optional| optional.child,
+        .optional => |optional| optional.child,
         else => T,
     };
 
     if (V == []const u8 or V == [:0]const u8) return arg_value;
-    if (@typeInfo(V) == .Int) return parse_value_int(V, arg_name, arg_value);
-    if (@typeInfo(V) == .Enum) return parse_value_enum(V, arg_name, arg_value);
+    if (@typeInfo(V) == .int) return parse_value_int(V, arg_name, arg_value);
+    if (@typeInfo(V) == .@"enum") return parse_value_enum(V, arg_name, arg_value);
     comptime unreachable;
 }
 
@@ -123,7 +132,7 @@ fn parse_value_int(comptime T: type, arg_name: []const u8, val: []const u8) T {
         switch (err) {
             error.Overflow => fatal(
                 "{s}: value exceeds {d}-bit {s} integer: '{s}'",
-                .{ arg_name, @typeInfo(T).Int.bits, @tagName(@typeInfo(T).Int.signedness), val },
+                .{ arg_name, @typeInfo(T).int.bits, @tagName(@typeInfo(T).int.signedness), val },
             ),
             error.InvalidCharacter => fatal(
                 "{s}: expected an integer value, but found '{s}' (invalid digit)",
@@ -151,7 +160,7 @@ test parse_value_int {
 /// assert(parse_value_enum(E, "test-enum", "not_ok"), .not_ok);
 /// ```
 fn parse_value_enum(comptime E: type, arg_name: []const u8, val: []const u8) E {
-    comptime assert(@typeInfo(E).Enum.is_exhaustive);
+    comptime assert(@typeInfo(E).@"enum".is_exhaustive);
 
     return std.meta.stringToEnum(E, val) orelse fatal(
         "{s}: expected one of {s}, but found '{s}'",
diff --git a/src/structs.zig b/src/structs.zig
index 55dd47d..3fd9a12 100644
--- a/src/structs.zig
+++ b/src/structs.zig
@@ -2,12 +2,12 @@ const std = @import("std");
 const StructField = std.builtin.Type.StructField;
 const assert = std.debug.assert;
 
-/// This is essentially `field.default_value`, but with a useful type instead of
+/// This is essentially `field.default_value_ptr`, but with a useful type instead of
 /// `?*const anyopaque`.
 pub fn default_value(comptime field: StructField) ?field.type {
     var out: ?field.type = null;
 
-    if (field.default_value) |default_opaque| {
+    if (field.default_value_ptr) |default_opaque| {
         out = @as(*const field.type, @ptrCast(@alignCast(default_opaque))).*;
     }
 
@@ -22,22 +22,22 @@ pub fn default_value(comptime field: StructField) ?field.type {
 /// Each field is of type `Data` and has the provided default, which may be
 /// undefined.
 pub fn struct_field_struct(comptime S: type, comptime Data: type, comptime default: ?Data) type {
-    assert(@typeInfo(S) == .Struct);
+    assert(@typeInfo(S) == .@"struct");
 
-    const fields_in = @typeInfo(S).Struct.fields;
+    const fields_in = @typeInfo(S).@"struct".fields;
     var fields_out: [fields_in.len]StructField = undefined;
 
     for (&fields_out, fields_in) |*field_out, field_in| {
         field_out.* = .{
             .name = field_in.name,
             .type = Data,
-            .default_value = if (default) |d| @as(?*const anyopaque, @ptrCast(&d)) else null,
+            .default_value_ptr = if (default) |d| @as(?*const anyopaque, @ptrCast(&d)) else null,
             .is_comptime = false,
             .alignment = if (@sizeOf(Data) > 0) @alignOf(Data) else 0,
         };
     }
 
-    return @Type(.{ .Struct = .{
+    return @Type(.{ .@"struct" = .{
         .layout = .auto,
         .fields = &fields_out,
         .decls = &.{},
diff --git a/src/zli.zig b/src/zli.zig
index 7c68dd3..95e0ecb 100644
--- a/src/zli.zig
+++ b/src/zli.zig
@@ -114,14 +114,14 @@ pub fn parse(args: *ArgIterator, comptime CLIArgs: type) CLIArgs {
     assert(args.skip()); // Discard executable name.
 
     return switch (@typeInfo(CLIArgs)) {
-        .Union => parse_commands(args, CLIArgs),
-        .Struct => parse_args(args, CLIArgs),
+        .@"union" => parse_commands(args, CLIArgs),
+        .@"struct" => parse_args(args, CLIArgs),
         else => unreachable,
     };
 }
 
 fn parse_commands(args: *ArgIterator, comptime Commands: type) Commands {
-    comptime assert(@typeInfo(Commands) == .Union);
+    comptime assert(@typeInfo(Commands) == .@"union");
     comptime assert(std.meta.fields(Commands).len > 1);
 
     const command = args.next() orelse fatal(
@@ -154,7 +154,7 @@ fn parse_args(args: *ArgIterator, comptime Args: type) Args {
         return {};
     }
 
-    comptime assert(@typeInfo(Args) == .Struct);
+    comptime assert(@typeInfo(Args) == .@"struct");
 
     comptime var fields: [std.meta.fields(Args).len]StructField = undefined;
     comptime var field_count = 0;
@@ -163,21 +163,26 @@ fn parse_args(args: *ArgIterator, comptime Args: type) Args {
 
     comptime for (std.meta.fields(Args)) |field| {
         if (strings.eql(field.name, "positional")) {
-            assert(@typeInfo(field.type) == .Struct);
+            assert(@typeInfo(field.type) == .@"struct");
 
             positional_fields = std.meta.fields(field.type);
 
             for (positional_fields) |positional_field| {
-                assert(structs.default_value(positional_field) == null);
-                argx.assert_valid_value_type(positional_field.type);
+                switch (@typeInfo(positional_field.type)) {
+                    .optional => |optional| {
+                        // if no default: will be required
+                        argx.assert_valid_value_type(optional.child);
+                    },
+                    else => argx.assert_valid_value_type(positional_field.type),
+                }
             }
         } else {
             switch (@typeInfo(field.type)) {
-                .Bool => {
-                    assert(structs.default_value(field).? == false); // boolean flags should have a default
+                .bool => {
+                    assert(structs.default_value(field).? == false); // boolean flags should have a default of false
                 },
-                .Optional => |optional| {
-                    assert(structs.default_value(field).? == null); // optional flags should have a default
+                .optional => |optional| {
+                    assert(structs.default_value(field).? == null); // optional flags should have a default of null
                     argx.assert_valid_value_type(optional.child);
                 },
                 else => {
@@ -274,8 +279,13 @@ fn parse_args(args: *ArgIterator, comptime Args: type) Args {
     if (@hasField(Args, "positional")) {
         assert(counts.positional <= positional_fields.len);
         inline for (positional_fields, 0..) |field, idx| {
-            if (counts.positional == idx) {
-                fatal("{s}: argument is required", .{field.name});
+            if (idx >= counts.positional) {
+                if (positional_fields[idx].default_value_ptr) |_| {
+                    // fill with default
+                    @field(result.positional, positional_fields[idx].name) = positional_fields[idx].defaultValue() orelse null;
+                } else {
+                    fatal("{s}: argument is required", .{field.name});
+                }
             }
         }
     }