diff --git a/reconcile-js/src/core.ts b/reconcile-js/src/core.ts
new file mode 100644
index 0000000..3dc78db
--- /dev/null
+++ b/reconcile-js/src/core.ts
@@ -0,0 +1,390 @@
+// Shared, platform-agnostic wrapper around the generated wasm-bindgen surface.
+//
+// The actual wasm bindings are injected by a platform-specific entrypoint:
+//   - `index.ts`     (web/node) instantiates the real WebAssembly module lazily
+//                    on first use via `initSync`.
+//   - `index.rn.ts`  (React Native / Hermes) links a wasm2js (pure-JS)
+//                    implementation, since Hermes does not expose a runtime
+//                    `WebAssembly` global. See `scripts/build-rn.mjs`.
+
+type WasmModule = typeof import('reconcile-text');
+
+/**
+ * The generated wasm-bindgen surface this library wraps, plus a hook to make
+ * sure the underlying module is ready. Supplied by a platform entrypoint.
+ */
+export interface WasmBackend {
+  CursorPosition: WasmModule['CursorPosition'];
+  TextWithCursors: WasmModule['TextWithCursors'];
+  reconcile: WasmModule['reconcile'];
+  reconcileWithHistory: WasmModule['reconcileWithHistory'];
+  diff: WasmModule['diff'];
+  undiff: WasmModule['undiff'];
+  /**
+   * Make the wasm module ready for use. Invoked before every operation, so it
+   * must be cheap and idempotent (a no-op once initialised).
+   */
+  ensureReady(): void;
+}
+
+// Define the enum values as const arrays to avoid duplication
+const BUILTIN_TOKENIZERS = ['Character', 'Line', 'Markdown', 'Word'] as const;
+
+/**
+ * Tokenisation strategies for text merging.
+ *
+ * These correspond to the built-in tokenizers available in the underlying WASM module.
+ */
+export type BuiltinTokenizer = (typeof BUILTIN_TOKENIZERS)[number];
+
+/**
+ * History classification for text spans in merge results.
+ *
+ * Indicates the origin of each text span in the merged document.
+ */
+export type History =
+  | 'Unchanged'
+  | 'AddedFromLeft'
+  | 'AddedFromRight'
+  | 'RemovedFromLeft'
+  | 'RemovedFromRight';
+
+/**
+ * Represents a text document with associated cursor positions.
+ *
+ * This interface is used both as input to reconcile functions (to specify where
+ * cursors are positioned in the original documents) and as output (with cursors
+ * automatically repositioned after merging).
+ */
+export interface TextWithCursors {
+  /** The document's entire content as a string */
+  text: string;
+
+  /**
+   * Array of cursor positions within the text. Can be empty if there are no cursors to track.
+   * Each cursor has a unique ID and position.
+   */
+  cursors: CursorPosition[];
+}
+
+/**
+ * Like `TextWithCursors`, but cursors may be null or undefined (treated as empty).
+ * Used as input where cursor tracking is optional.
+ */
+export interface TextWithOptionalCursors {
+  /** The document's entire content as a string */
+  text: string;
+
+  /**
+   * Array of cursor positions within the text. Can be null, undefined, or empty
+   * if there are no cursors to track. Each cursor has a unique ID and position.
+   */
+  cursors: null | undefined | CursorPosition[];
+}
+
+/**
+ * Represents a cursor position within a text document.
+ *
+ * Cursors are automatically repositioned during text merging to maintain their
+ * relative positions as text is inserted, deleted, or modified around them.
+ */
+export interface CursorPosition {
+  /** Unique identifier for the cursor (can be any number, must be unique within the document) */
+  id: number;
+
+  /** Character position in the text, 0-based index from the beginning of the document */
+  position: number;
+}
+
+/**
+ * Represents a merged text document with cursor positions and detailed change history.
+ *
+ * This is the return type of `reconcileWithHistory()` and provides complete information
+ * about how the merge was performed, including which parts of the final text came from
+ * which source documents.
+ */
+export interface TextWithCursorsAndHistory {
+  /** The merged document's entire content */
+  text: string;
+
+  /**
+   * Array of cursor positions within the merged text. Can be empty if there are no cursors to track.
+   * All cursors are automatically repositioned from the left and right documents.
+   */
+  cursors: CursorPosition[];
+
+  /**
+   * Detailed provenance information showing the origin of each text span in the result.
+   * Each span indicates whether it was unchanged, added from left, added from right, etc.
+   */
+  history: SpanWithHistory[];
+}
+
+/**
+ * Represents a span of text in the merged result with its change history.
+ *
+ * This shows exactly which source document contributed each piece of text to the
+ * final merged result. Useful for understanding merge decisions and creating
+ * visualisations of how documents were combined.
+ */
+export interface SpanWithHistory {
+  /** The text content of this span */
+  text: string;
+
+  /** The origin of this text span in the merge result */
+  history: History;
+}
+
+/** The public, synchronous API surface, identical across platforms. */
+export interface ReconcileApi {
+  /**
+   * Merges three versions of text using intelligent conflict resolution.
+   *
+   * This is the primary function for 3-way text merging. Unlike traditional merge tools
+   * that produce conflict markers, this function automatically resolves conflicts by
+   * applying both sets of changes where possible.
+   *
+   * @param original - The original/base version of the text that both sides diverged from
+   * @param left - The left version of the text (either string or TextWithCursors with cursor positions)
+   * @param right - The right version of the text (either string or TextWithCursors with cursor positions)
+   * @param tokenizer - The tokenisation strategy: "Word" (default, recommended for prose),
+   *                    "Character" (fine-grained), or "Line" (similar to git merge)
+   * @returns The reconciled text with automatically repositioned cursor positions
+   *
+   * @example
+   * ```typescript
+   * const original = "Hello world";
+   * const left = "Hello beautiful world";    // Added "beautiful"
+   * const right = "Hi world";                // Changed "Hello" to "Hi"
+   *
+   * const result = reconcile(original, left, right);
+   * console.log(result.text); // "Hi beautiful world"
+   * ```
+   */
+  reconcile(
+    original: string,
+    left: string | TextWithOptionalCursors,
+    right: string | TextWithOptionalCursors,
+    tokenizer?: BuiltinTokenizer
+  ): TextWithCursors;
+
+  /**
+   * Generates a compact diff representation between an original and changed text.
+   *
+   * These can be parsed and unpacked using the `undiff` function or the Rust crate's EditedText::from_diff.
+   * Cursor positions are omitted from the diff result.
+   *
+   * This function computes the differences between two versions of text and returns
+   * a compact representation of those changes.
+   *
+   * @param original - The original/base version of the text
+   * @param changed - The modified version of the text (either string or TextWithCursors with cursor positions)
+   * @param tokenizer - The tokenisation strategy, which is the same as used in `reconcile`.
+   * @returns An array of inserts (strings), deletes (negative integers), and retained spans (positive integers).
+   */
+  diff(
+    original: string,
+    changed: string | TextWithOptionalCursors,
+    tokenizer?: BuiltinTokenizer
+  ): Array<number | string>;
+
+  /**
+   * Applies a compact diff to an original text to reconstruct the changed version.
+   *
+   * This function takes an original text and a compact diff representation (as produced
+   * by the `diff` function) and reconstructs the modified text.
+   *
+   * @param original - The original/base version of the text
+   * @param diff - The compact diff array (inserts as strings, deletes as negative integers, retained spans as positive integers)
+   * @param tokenizer - The tokenisation strategy, which is the same as used in `reconcile`.
+   * @returns The reconstructed changed text as a string.
+   */
+  undiff(
+    original: string,
+    diff: Array<number | bigint | string>,
+    tokenizer?: BuiltinTokenizer
+  ): string;
+
+  /**
+   * Merges three versions of text and returns detailed provenance information.
+   *
+   * This function behaves like `reconcile()` but also provides
+   * detailed historical information about the origin of each text span in the result.
+   * This is valuable for understanding how the merge was performed and which changes
+   * came from which source.
+   *
+   * Note: Computing the history is computationally more expensive than the basic merge.
+   *
+   * @param original - The original/base version of the text that both sides diverged from
+   * @param left - The left version of the text (either string or TextWithCursors with cursor positions)
+   * @param right - The right version of the text (either string or TextWithCursors with cursor positions)
+   * @param tokenizer - The tokenisation strategy: "Word" (default, recommended for prose),
+   *                    "Character" (fine-grained), or "Line" (similar to git merge)
+   * @returns The reconciled text with cursor positions and detailed change history
+   *
+   * @example
+   * ```typescript
+   * const original = "Hello world";
+   * const left = "Hello beautiful world";
+   * const right = "Hi world";
+   *
+   * const result = reconcileWithHistory(original, left, right);
+   * console.log(result.text); // "Hi beautiful world"
+   * console.log(result.history); // Array of SpanWithHistory objects showing change origins
+   * ```
+   */
+  reconcileWithHistory(
+    original: string,
+    left: string | TextWithOptionalCursors,
+    right: string | TextWithOptionalCursors,
+    tokenizer?: BuiltinTokenizer
+  ): TextWithCursorsAndHistory;
+}
+
+const UNSUPPORTED_TOKENIZER_ERROR = `Unsupported tokenizer, only ${BUILTIN_TOKENIZERS.join(
+  ', '
+)} are supported`;
+
+/**
+ * Build the public {@link ReconcileApi} on top of a {@link WasmBackend}.
+ *
+ * Each operation calls `backend.ensureReady()` first, then marshals JS values
+ * into the wasm representation, invokes the binding, and frees the wasm-side
+ * objects. The behaviour is identical regardless of whether the backend is a
+ * real WebAssembly module or its wasm2js translation.
+ */
+export function makeReconcileApi(backend: WasmBackend): ReconcileApi {
+  function assertTokenizer(tokenizer: BuiltinTokenizer): void {
+    if (!BUILTIN_TOKENIZERS.includes(tokenizer)) {
+      throw new Error(UNSUPPORTED_TOKENIZER_ERROR);
+    }
+  }
+
+  function toWasmTextWithCursors(text: string | TextWithOptionalCursors) {
+    const isInputString = typeof text === 'string';
+    const innerText = isInputString ? text : text.text;
+    const innerCursors = isInputString ? [] : (text.cursors ?? []);
+
+    return new backend.TextWithCursors(
+      innerText,
+      innerCursors.map(({ id, position }) => new backend.CursorPosition(id, position))
+    );
+  }
+
+  function toTextWithCursors(textWithCursor: {
+    text(): string;
+    cursors(): Array<{ id(): number; characterIndex(): number; free(): void }>;
+  }): TextWithCursors {
+    const wasmCursors = textWithCursor.cursors();
+    const cursors = wasmCursors.map((cursor) => ({
+      id: cursor.id(),
+      position: cursor.characterIndex(),
+    }));
+    for (const cursor of wasmCursors) {
+      cursor.free();
+    }
+
+    return {
+      text: textWithCursor.text(),
+      cursors,
+    };
+  }
+
+  function toSpanWithHistory(span: {
+    text(): string;
+    history(): History;
+    free(): void;
+  }): SpanWithHistory {
+    const result = {
+      text: span.text(),
+      history: span.history(),
+    };
+    span.free();
+    return result;
+  }
+
+  function reconcile(
+    original: string,
+    left: string | TextWithOptionalCursors,
+    right: string | TextWithOptionalCursors,
+    tokenizer: BuiltinTokenizer = 'Word'
+  ): TextWithCursors {
+    backend.ensureReady();
+    assertTokenizer(tokenizer);
+
+    const leftCursor = toWasmTextWithCursors(left);
+    const rightCursor = toWasmTextWithCursors(right);
+
+    const result = backend.reconcile(original, leftCursor, rightCursor, tokenizer);
+
+    leftCursor.free();
+    rightCursor.free();
+
+    const jsResult = toTextWithCursors(result);
+    result.free();
+
+    return jsResult;
+  }
+
+  function diff(
+    original: string,
+    changed: string | TextWithOptionalCursors,
+    tokenizer: BuiltinTokenizer = 'Word'
+  ): Array<number | string> {
+    backend.ensureReady();
+    assertTokenizer(tokenizer);
+
+    const changedWasm = toWasmTextWithCursors(changed);
+
+    const result = backend.diff(original, changedWasm, tokenizer);
+
+    changedWasm.free();
+
+    return result.map((item) => (typeof item === 'bigint' ? Number(item) : item));
+  }
+
+  function undiff(
+    original: string,
+    diffValue: Array<number | bigint | string>,
+    tokenizer: BuiltinTokenizer = 'Word'
+  ): string {
+    backend.ensureReady();
+    assertTokenizer(tokenizer);
+
+    return backend.undiff(original, diffValue, tokenizer);
+  }
+
+  function reconcileWithHistory(
+    original: string,
+    left: string | TextWithOptionalCursors,
+    right: string | TextWithOptionalCursors,
+    tokenizer: BuiltinTokenizer = 'Word'
+  ): TextWithCursorsAndHistory {
+    backend.ensureReady();
+    assertTokenizer(tokenizer);
+
+    const leftCursor = toWasmTextWithCursors(left);
+    const rightCursor = toWasmTextWithCursors(right);
+
+    const result = backend.reconcileWithHistory(
+      original,
+      leftCursor,
+      rightCursor,
+      tokenizer
+    );
+
+    leftCursor.free();
+    rightCursor.free();
+
+    const jsResult = toTextWithCursors(result);
+    const history = result.history().map(toSpanWithHistory);
+    result.free();
+
+    return {
+      ...jsResult,
+      history,
+    };
+  }
+
+  return { reconcile, diff, undiff, reconcileWithHistory };
+}