@oomkapwn/enquire-mcp 2.15.0 → 2.17.0

package/dist/hnsw.d.ts

@@ -6,6 +6,42 @@ export interface LabeledVector {
     /** L2-normalized vector. Caller is responsible for the normalization. */
     vector: Float32Array;
 }
+/**
+ * v2.16.0 — sidecar metadata persisted alongside the .hnsw.bin index. Used
+ * for staleness detection on boot: if `signature` matches the current
+ * embed-db's signature (computed by `EmbedDb.computeSignature()`), the
+ * pre-built HNSW is loaded; otherwise it's rebuilt from scratch.
+ *
+ * Stored as JSON next to the binary index (`<file>.meta.json`). Keep this
+ * format stable — bumping `formatVersion` invalidates all on-disk
+ * indexes for users on the new version (they'll rebuild on next boot,
+ * which is harmless but visible).
+ */
+export interface HnswPersistedMeta {
+    formatVersion: 1;
+    /** Embedder dim — must match the corpus the index will be queried with. */
+    dim: number;
+    /** Vector count at write time. */
+    size: number;
+    /**
+     * Embed-db signature at write time — when this differs from the current
+     * embed-db's signature, the persisted index is stale and should be
+     * rebuilt. We use rowcount + max-id + dim as a tractable signature
+     * (full content-hash would require reading every vector).
+     */
+    signature: string;
+    /** Row label → source row map needed to reconstruct hits. JSON-friendly. */
+    rowsByLabel: Record<string, {
+        rel_path: string;
+        chunk_index: number;
+        line_start: number;
+        line_end: number;
+        text_preview: string;
+        kind: "md" | "pdf";
+    }>;
+    /** ISO timestamp of the write — informational. */
+    writtenAt: string;
+}
 /** Build-time HNSW parameters. Defaults tuned for 384-dim cosine on PKM data. */
 export interface HnswBuildOptions {
     /** Embedding dimensionality (must match the corpus). */
@@ -52,6 +88,25 @@ export interface HnswIndex {
         labels: number[];
         distances: number[];
     };
+    /**
+     * v2.16.0 — persist the index to disk for fast reload on next serve
+     * start. Writes the binary index to `<file>.bin` and a JSON meta
+     * sidecar to `<file>.meta.json` containing the embed-db signature,
+     * dim, size, and label→row map. Returns true on successful write.
+     *
+     * Caller is responsible for choosing `file` (typically alongside the
+     * embed-db with `.hnsw` suffix). We separate binary + meta files so
+     * a partial write (e.g. crash mid-flush) leaves the meta missing,
+     * which the loader treats as "no usable index" → rebuild from scratch.
+     */
+    saveTo(file: string, rowsByLabel: ReadonlyMap<number, {
+        rel_path: string;
+        chunk_index: number;
+        line_start: number;
+        line_end: number;
+        text_preview: string;
+        kind: "md" | "pdf";
+    }>, signature: string): Promise<boolean>;
 }
 /**
  * Build a fresh in-memory HNSW from labeled vectors.
@@ -65,6 +120,23 @@ export interface HnswIndex {
  * is less than the input count, or if `hnswlib-wasm` failed to load.
  */
 export declare function buildHnsw(vectors: ReadonlyArray<LabeledVector>, opts: HnswBuildOptions): Promise<HnswIndex>;
+/**
+ * v2.16.0 — load a previously-persisted HNSW index from disk. Returns
+ * `null` (with a stderr warning) if:
+ *   • Either the .bin or .meta.json file is missing
+ *   • The meta's `signature` doesn't match the caller's current signature
+ *   • The meta's `formatVersion` doesn't match
+ *   • The native lib fails to load the .bin (corrupt / dim mismatch)
+ *
+ * On success returns `{ index, rowsByLabel }` so the caller can wire
+ * both into `searchHybrid`'s `hnsw` context without rebuilding from
+ * scratch. Typical boot-time win: ~25s rebuild → ~50ms load on a
+ * 50K-chunk vault.
+ */
+export declare function loadHnswFromDisk(file: string, expectedSignature: string): Promise<{
+    index: HnswIndex;
+    rowsByLabel: Map<number, HnswPersistedMeta["rowsByLabel"][string]>;
+} | null>;
 /**
  * Convert HNSW search results to EmbedSearchHit using a label → source-row
  * lookup. The label was assigned by the caller at build time (typically

package/dist/hnsw.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"hnsw.d.ts","sourceRoot":"","sources":["../src/hnsw.ts"],"names":[],"mappings":"AA0CA,OAAO,KAAK,EAAE,cAAc,EAAE,MAAM,eAAe,CAAC;AAEpD,4DAA4D;AAC5D,MAAM,WAAW,aAAa;IAC5B,wFAAwF;IACxF,KAAK,EAAE,MAAM,CAAC;IACd,yEAAyE;IACzE,MAAM,EAAE,YAAY,CAAC;CACtB;AAED,iFAAiF;AACjF,MAAM,WAAW,gBAAgB;IAC/B,wDAAwD;IACxD,GAAG,EAAE,MAAM,CAAC;IACZ,8EAA8E;IAC9E,WAAW,EAAE,MAAM,CAAC;IACpB;;;OAGG;IACH,CAAC,CAAC,EAAE,MAAM,CAAC;IACX;;;OAGG;IACH,cAAc,CAAC,EAAE,MAAM,CAAC;IACxB,oEAAoE;IACpE,IAAI,CAAC,EAAE,MAAM,CAAC;CACf;AAED,4BAA4B;AAC5B,MAAM,WAAW,gBAAgB;IAC/B;;;OAGG;IACH,EAAE,CAAC,EAAE,MAAM,CAAC;CACb;AAED;;;;GAIG;AACH,MAAM,WAAW,SAAS;IACxB,6BAA6B;IAC7B,QAAQ,CAAC,GAAG,EAAE,MAAM,CAAC;IACrB,+CAA+C;IAC/C,QAAQ,CAAC,IAAI,EAAE,MAAM,CAAC;IACtB;;;;OAIG;IACH,SAAS,CAAC,QAAQ,EAAE,YAAY,EAAE,CAAC,EAAE,MAAM,EAAE,IAAI,CAAC,EAAE,gBAAgB,GAAG;QAAE,MAAM,EAAE,MAAM,EAAE,CAAC;QAAC,SAAS,EAAE,MAAM,EAAE,CAAA;KAAE,CAAC;CAClH;AAgDD;;;;;;;;;;GAUG;AACH,wBAAsB,SAAS,CAAC,OAAO,EAAE,aAAa,CAAC,aAAa,CAAC,EAAE,IAAI,EAAE,gBAAgB,GAAG,OAAO,CAAC,SAAS,CAAC,CAmDjH;AAED;;;;;;;GAOG;AACH,wBAAgB,iBAAiB,CAC/B,MAAM,EAAE;IAAE,MAAM,EAAE,MAAM,EAAE,CAAC;IAAC,SAAS,EAAE,MAAM,EAAE,CAAA;CAAE,EACjD,UAAU,EAAE,WAAW,CACrB,MAAM,EACN;IACE,QAAQ,EAAE,MAAM,CAAC;IACjB,WAAW,EAAE,MAAM,CAAC;IACpB,UAAU,EAAE,MAAM,CAAC;IACnB,QAAQ,EAAE,MAAM,CAAC;IACjB,YAAY,EAAE,MAAM,CAAC;IACrB,IAAI,EAAE,IAAI,GAAG,KAAK,CAAC;CACpB,CACF,GACA,cAAc,EAAE,CAsBlB"}
+{"version":3,"file":"hnsw.d.ts","sourceRoot":"","sources":["../src/hnsw.ts"],"names":[],"mappings":"AA0CA,OAAO,KAAK,EAAE,cAAc,EAAE,MAAM,eAAe,CAAC;AAEpD,4DAA4D;AAC5D,MAAM,WAAW,aAAa;IAC5B,wFAAwF;IACxF,KAAK,EAAE,MAAM,CAAC;IACd,yEAAyE;IACzE,MAAM,EAAE,YAAY,CAAC;CACtB;AAED;;;;;;;;;;GAUG;AACH,MAAM,WAAW,iBAAiB;IAChC,aAAa,EAAE,CAAC,CAAC;IACjB,2EAA2E;IAC3E,GAAG,EAAE,MAAM,CAAC;IACZ,kCAAkC;IAClC,IAAI,EAAE,MAAM,CAAC;IACb;;;;;OAKG;IACH,SAAS,EAAE,MAAM,CAAC;IAClB,4EAA4E;IAC5E,WAAW,EAAE,MAAM,CACjB,MAAM,EACN;QACE,QAAQ,EAAE,MAAM,CAAC;QACjB,WAAW,EAAE,MAAM,CAAC;QACpB,UAAU,EAAE,MAAM,CAAC;QACnB,QAAQ,EAAE,MAAM,CAAC;QACjB,YAAY,EAAE,MAAM,CAAC;QACrB,IAAI,EAAE,IAAI,GAAG,KAAK,CAAC;KACpB,CACF,CAAC;IACF,kDAAkD;IAClD,SAAS,EAAE,MAAM,CAAC;CACnB;AAED,iFAAiF;AACjF,MAAM,WAAW,gBAAgB;IAC/B,wDAAwD;IACxD,GAAG,EAAE,MAAM,CAAC;IACZ,8EAA8E;IAC9E,WAAW,EAAE,MAAM,CAAC;IACpB;;;OAGG;IACH,CAAC,CAAC,EAAE,MAAM,CAAC;IACX;;;OAGG;IACH,cAAc,CAAC,EAAE,MAAM,CAAC;IACxB,oEAAoE;IACpE,IAAI,CAAC,EAAE,MAAM,CAAC;CACf;AAED,4BAA4B;AAC5B,MAAM,WAAW,gBAAgB;IAC/B;;;OAGG;IACH,EAAE,CAAC,EAAE,MAAM,CAAC;CACb;AAED;;;;GAIG;AACH,MAAM,WAAW,SAAS;IACxB,6BAA6B;IAC7B,QAAQ,CAAC,GAAG,EAAE,MAAM,CAAC;IACrB,+CAA+C;IAC/C,QAAQ,CAAC,IAAI,EAAE,MAAM,CAAC;IACtB;;;;OAIG;IACH,SAAS,CAAC,QAAQ,EAAE,YAAY,EAAE,CAAC,EAAE,MAAM,EAAE,IAAI,CAAC,EAAE,gBAAgB,GAAG;QAAE,MAAM,EAAE,MAAM,EAAE,CAAC;QAAC,SAAS,EAAE,MAAM,EAAE,CAAA;KAAE,CAAC;IACjH;;;;;;;;;;OAUG;IACH,MAAM,CACJ,IAAI,EAAE,MAAM,EACZ,WAAW,EAAE,WAAW,CACtB,MAAM,EACN;QACE,QAAQ,EAAE,MAAM,CAAC;QACjB,WAAW,EAAE,MAAM,CAAC;QACpB,UAAU,EAAE,MAAM,CAAC;QACnB,QAAQ,EAAE,MAAM,CAAC;QACjB,YAAY,EAAE,MAAM,CAAC;QACrB,IAAI,EAAE,IAAI,GAAG,KAAK,CAAC;KACpB,CACF,EACD,SAAS,EAAE,MAAM,GAChB,OAAO,CAAC,OAAO,CAAC,CAAC;CACrB;AAmDD;;;;;;;;;;GAUG;AACH,wBAAsB,SAAS,CAAC,OAAO,EAAE,aAAa,CAAC,aAAa,CAAC,EAAE,IAAI,EAAE,gBAAgB,GAAG,OAAO,CAAC,SAAS,CAAC,CAqCjH;AAgDD;;;;;;;;;;;;GAYG;AACH,wBAAsB,gBAAgB,CACpC,IAAI,EAAE,MAAM,EACZ,iBAAiB,EAAE,MAAM,GACxB,OAAO,CAAC;IAAE,KAAK,EAAE,SAAS,CAAC;IAAC,WAAW,EAAE,GAAG,CAAC,MAAM,EAAE,iBAAiB,CAAC,aAAa,CAAC,CAAC,MAAM,CAAC,CAAC,CAAA;CAAE,GAAG,IAAI,CAAC,CAwD1G;AAED;;;;;;;GAOG;AACH,wBAAgB,iBAAiB,CAC/B,MAAM,EAAE;IAAE,MAAM,EAAE,MAAM,EAAE,CAAC;IAAC,SAAS,EAAE,MAAM,EAAE,CAAA;CAAE,EACjD,UAAU,EAAE,WAAW,CACrB,MAAM,EACN;IACE,QAAQ,EAAE,MAAM,CAAC;IACjB,WAAW,EAAE,MAAM,CAAC;IACpB,UAAU,EAAE,MAAM,CAAC;IACnB,QAAQ,EAAE,MAAM,CAAC;IACjB,YAAY,EAAE,MAAM,CAAC;IACrB,IAAI,EAAE,IAAI,GAAG,KAAK,CAAC;CACpB,CACF,GACA,cAAc,EAAE,CAsBlB"}

package/dist/hnsw.js

@@ -104,9 +104,17 @@ export async function buildHnsw(vectors, opts) {
         // an intermediate; we use a plain spread which is fast and explicit.
         ctor.addPoint(Array.from(v.vector), v.label);
     }
+    return wrapNativeIndex(ctor, dim, vectors.length);
+}
+/**
+ * v2.16.0 — wrap a native hnswlib-node index (built fresh OR loaded from
+ * disk) as our `HnswIndex` type. Factored out of `buildHnsw` so the
+ * load-from-disk path returns the same shape without re-running addPoint.
+ */
+function wrapNativeIndex(ctor, dim, size) {
     return {
         dim,
-        size: vectors.length,
+        size,
         searchKnn(queryVec, k, qOpts) {
             if (queryVec.length !== dim) {
                 throw new Error(`HnswIndex.searchKnn: query dim ${queryVec.length} ≠ index dim ${dim}`);
@@ -117,9 +125,98 @@ export async function buildHnsw(vectors, opts) {
             ctor.setEf(ef);
             const result = ctor.searchKnn(Array.from(queryVec), k, undefined);
             return { labels: result.neighbors, distances: result.distances };
+        },
+        async saveTo(file, rowsByLabel, signature) {
+            const fs = await import("node:fs/promises");
+            const path = await import("node:path");
+            // Write the binary index to <file>.bin and the JSON meta sidecar
+            // to <file>.meta.json. We separate them so a partial-write (e.g.
+            // crash mid-flush) leaves meta missing → loader rebuilds.
+            await fs.mkdir(path.dirname(file), { recursive: true });
+            const binFile = `${file}.bin`;
+            const metaFile = `${file}.meta.json`;
+            // hnswlib-node's writeIndex writes to a host filesystem path
+            // directly — much simpler than the WASM Emscripten FS plumbing.
+            await ctor.writeIndex(binFile);
+            const meta = {
+                formatVersion: 1,
+                dim,
+                size,
+                signature,
+                rowsByLabel: Object.fromEntries(rowsByLabel),
+                writtenAt: new Date().toISOString()
+            };
+            await fs.writeFile(metaFile, JSON.stringify(meta, null, 2), "utf8");
+            return true;
         }
     };
 }
+/**
+ * v2.16.0 — load a previously-persisted HNSW index from disk. Returns
+ * `null` (with a stderr warning) if:
+ *   • Either the .bin or .meta.json file is missing
+ *   • The meta's `signature` doesn't match the caller's current signature
+ *   • The meta's `formatVersion` doesn't match
+ *   • The native lib fails to load the .bin (corrupt / dim mismatch)
+ *
+ * On success returns `{ index, rowsByLabel }` so the caller can wire
+ * both into `searchHybrid`'s `hnsw` context without rebuilding from
+ * scratch. Typical boot-time win: ~25s rebuild → ~50ms load on a
+ * 50K-chunk vault.
+ */
+export async function loadHnswFromDisk(file, expectedSignature) {
+    const fs = await import("node:fs/promises");
+    const binFile = `${file}.bin`;
+    const metaFile = `${file}.meta.json`;
+    let metaRaw;
+    try {
+        metaRaw = await fs.readFile(metaFile, "utf8");
+    }
+    catch {
+        return null; // No meta → no persisted index (or partial write).
+    }
+    let meta;
+    try {
+        meta = JSON.parse(metaRaw);
+    }
+    catch (err) {
+        process.stderr.write(`enquire: HNSW meta at ${metaFile} is malformed; rebuilding — ${err instanceof Error ? err.message : String(err)}\n`);
+        return null;
+    }
+    if (meta.formatVersion !== 1) {
+        process.stderr.write(`enquire: HNSW meta format ${meta.formatVersion} ≠ expected 1; rebuilding (this happens on enquire-mcp upgrade)\n`);
+        return null;
+    }
+    if (meta.signature !== expectedSignature) {
+        process.stderr.write(`enquire: HNSW persisted index is stale (signature mismatch — embed-db changed since last write); rebuilding\n`);
+        return null;
+    }
+    // Bin file present?
+    try {
+        await fs.access(binFile);
+    }
+    catch {
+        process.stderr.write(`enquire: HNSW meta exists but ${binFile} is missing; rebuilding\n`);
+        return null;
+    }
+    // Load the native binary.
+    const lib = await loadHnswlib();
+    const ctor = new lib.HierarchicalNSW("cosine", meta.dim);
+    try {
+        await ctor.readIndex(binFile);
+    }
+    catch (err) {
+        process.stderr.write(`enquire: HNSW readIndex failed at ${binFile}; rebuilding — ${err instanceof Error ? err.message : String(err)}\n`);
+        return null;
+    }
+    const index = wrapNativeIndex(ctor, meta.dim, meta.size);
+    // Reconstruct the row map.
+    const rowsByLabel = new Map();
+    for (const [labelStr, row] of Object.entries(meta.rowsByLabel)) {
+        rowsByLabel.set(Number.parseInt(labelStr, 10), row);
+    }
+    return { index, rowsByLabel };
+}
 /**
  * Convert HNSW search results to EmbedSearchHit using a label → source-row
  * lookup. The label was assigned by the caller at build time (typically

package/dist/hnsw.js.map

@@ -1 +1 @@
-{"version":3,"file":"hnsw.js","sourceRoot":"","sources":["../src/hnsw.ts"],"names":[],"mappings":"AAAA,0EAA0E;AAC1E,EAAE;AACF,4EAA4E;AAC5E,2EAA2E;AAC3E,2EAA2E;AAC3E,wEAAwE;AACxE,wEAAwE;AACxE,+EAA+E;AAC/E,EAAE;AACF,kDAAkD;AAClD,EAAE;AACF,sBAAsB;AACtB,wEAAwE;AACxE,yEAAyE;AACzE,oEAAoE;AACpE,wEAAwE;AACxE,uEAAuE;AACvE,2EAA2E;AAC3E,wEAAwE;AACxE,0EAA0E;AAC1E,EAAE;AACF,yEAAyE;AACzE,uEAAuE;AACvE,uEAAuE;AACvE,sEAAsE;AACtE,wEAAwE;AACxE,4DAA4D;AAC5D,EAAE;AACF,mEAAmE;AACnE,sEAAsE;AACtE,qEAAqE;AACrE,kEAAkE;AAClE,EAAE;AACF,iEAAiE;AACjE,wEAAwE;AACxE,oEAAoE;AACpE,EAAE;AACF,sEAAsE;AACtE,0EAA0E;AAC1E,iEAAiE;AACjE,WAAW;AAkFX,IAAI,YAAY,GAA6B,IAAI,CAAC;AAClD,KAAK,UAAU,WAAW;IACxB,IAAI,YAAY;QAAE,OAAO,YAAY,CAAC;IACtC,IAAI,CAAC;QACH,MAAM,GAAG,GAAG,CAAC,MAAM,MAAM,CAAC,cAAc,CAAC,CAAiE,CAAC;QAC3G,qEAAqE;QACrE,mDAAmD;QACnD,MAAM,GAAG,GAAG,GAAG,CAAC,OAAO,IAAK,GAAyB,CAAC;QACtD,IAAI,OAAO,GAAG,CAAC,eAAe,KAAK,UAAU,EAAE,CAAC;YAC9C,MAAM,IAAI,KAAK,CAAC,+DAA+D,CAAC,CAAC;QACnF,CAAC;QACD,YAAY,GAAG,GAAG,CAAC;QACnB,OAAO,YAAY,CAAC;IACtB,CAAC;IAAC,OAAO,GAAG,EAAE,CAAC;QACb,MAAM,GAAG,GAAG,GAAG,YAAY,KAAK,CAAC,CAAC,CAAC,GAAG,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,GAAG,CAAC,CAAC;QAC7D,MAAM,IAAI,KAAK,CACb,kFAAkF;YAChF,gGAAgG;YAChG,qBAAqB,GAAG,EAAE,CAC7B,CAAC;IACJ,CAAC;AACH,CAAC;AAED;;;;;;;;;;GAUG;AACH,MAAM,CAAC,KAAK,UAAU,SAAS,CAAC,OAAqC,EAAE,IAAsB;IAC3F,MAAM,GAAG,GAAG,IAAI,CAAC,GAAG,CAAC;IACrB,IAAI,OAAO,CAAC,MAAM,GAAG,IAAI,CAAC,WAAW,EAAE,CAAC;QACtC,MAAM,IAAI,KAAK,CACb,6BAA6B,OAAO,CAAC,MAAM,wBAAwB,IAAI,CAAC,WAAW,sBAAsB,CAC1G,CAAC;IACJ,CAAC;IACD,MAAM,CAAC,GAAG,IAAI,CAAC,CAAC,IAAI,EAAE,CAAC;IACvB,MAAM,cAAc,GAAG,IAAI,CAAC,cAAc,IAAI,GAAG,CAAC;IAClD,MAAM,IAAI,GAAG,IAAI,CAAC,IAAI,IAAI,GAAG,CAAC;IAE9B,gEAAgE;IAChE,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,OAAO,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;QACxC,MAAM,CAAC,GAAG,OAAO,CAAC,CAAC,CAAC,CAAC;QACrB,IAAI,CAAC,CAAC;YAAE,SAAS;QACjB,IAAI,CAAC,CAAC,MAAM,CAAC,MAAM,KAAK,GAAG,EAAE,CAAC;YAC5B,MAAM,IAAI,KAAK,CAAC,8BAA8B,CAAC,YAAY,CAAC,CAAC,MAAM,CAAC,MAAM,cAAc,GAAG,EAAE,CAAC,CAAC;QACjG,CAAC;IACH,CAAC;IAED,MAAM,GAAG,GAAG,MAAM,WAAW,EAAE,CAAC;IAChC,MAAM,IAAI,GAAG,IAAI,GAAG,CAAC,eAAe,CAAC,QAAQ,EAAE,GAAG,CAAC,CAAC;IACpD,wEAAwE;IACxE,yEAAyE;IACzE,uBAAuB;IACvB,IAAI,CAAC,SAAS,CAAC,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,WAAW,EAAE,CAAC,CAAC,EAAE,CAAC,EAAE,cAAc,EAAE,IAAI,CAAC,CAAC;IAEvE,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,OAAO,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;QACxC,MAAM,CAAC,GAAG,OAAO,CAAC,CAAC,CAAC,CAAC;QACrB,IAAI,CAAC,CAAC;YAAE,SAAS;QACjB,kEAAkE;QAClE,sEAAsE;QACtE,qEAAqE;QACrE,IAAI,CAAC,QAAQ,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,CAAC,MAAM,CAAC,EAAE,CAAC,CAAC,KAAK,CAAC,CAAC;IAC/C,CAAC;IAED,OAAO;QACL,GAAG;QACH,IAAI,EAAE,OAAO,CAAC,MAAM;QACpB,SAAS,CAAC,QAAsB,EAAE,CAAS,EAAE,KAAwB;YACnE,IAAI,QAAQ,CAAC,MAAM,KAAK,GAAG,EAAE,CAAC;gBAC5B,MAAM,IAAI,KAAK,CAAC,kCAAkC,QAAQ,CAAC,MAAM,gBAAgB,GAAG,EAAE,CAAC,CAAC;YAC1F,CAAC;YACD,oEAAoE;YACpE,0CAA0C;YAC1C,MAAM,EAAE,GAAG,IAAI,CAAC,GAAG,CAAC,KAAK,EAAE,EAAE,IAAI,GAAG,EAAE,CAAC,CAAC,CAAC;YACzC,IAAI,CAAC,KAAK,CAAC,EAAE,CAAC,CAAC;YACf,MAAM,MAAM,GAAG,IAAI,CAAC,SAAS,CAAC,KAAK,CAAC,IAAI,CAAC,QAAQ,CAAC,EAAE,CAAC,EAAE,SAAS,CAAC,CAAC;YAClE,OAAO,EAAE,MAAM,EAAE,MAAM,CAAC,SAAS,EAAE,SAAS,EAAE,MAAM,CAAC,SAAS,EAAE,CAAC;QACnE,CAAC;KACF,CAAC;AACJ,CAAC;AAED;;;;;;;GAOG;AACH,MAAM,UAAU,iBAAiB,CAC/B,MAAiD,EACjD,UAUC;IAED,MAAM,IAAI,GAAqB,EAAE,CAAC;IAClC,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,MAAM,CAAC,MAAM,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;QAC9C,MAAM,KAAK,GAAG,MAAM,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC;QAC/B,MAAM,QAAQ,GAAG,MAAM,CAAC,SAAS,CAAC,CAAC,CAAC,CAAC;QACrC,IAAI,KAAK,KAAK,SAAS,IAAI,QAAQ,KAAK,SAAS;YAAE,SAAS;QAC5D,MAAM,GAAG,GAAG,UAAU,CAAC,GAAG,CAAC,KAAK,CAAC,CAAC;QAClC,IAAI,CAAC,GAAG;YAAE,SAAS,CAAC,mDAAmD;QACvE,wDAAwD;QACxD,kEAAkE;QAClE,MAAM,KAAK,GAAG,CAAC,GAAG,QAAQ,CAAC;QAC3B,IAAI,CAAC,IAAI,CAAC;YACR,QAAQ,EAAE,GAAG,CAAC,QAAQ;YACtB,WAAW,EAAE,GAAG,CAAC,WAAW;YAC5B,UAAU,EAAE,GAAG,CAAC,UAAU;YAC1B,QAAQ,EAAE,GAAG,CAAC,QAAQ;YACtB,YAAY,EAAE,GAAG,CAAC,YAAY;YAC9B,KAAK;YACL,IAAI,EAAE,GAAG,CAAC,IAAI;SACf,CAAC,CAAC;IACL,CAAC;IACD,OAAO,IAAI,CAAC;AACd,CAAC"}
+{"version":3,"file":"hnsw.js","sourceRoot":"","sources":["../src/hnsw.ts"],"names":[],"mappings":"AAAA,0EAA0E;AAC1E,EAAE;AACF,4EAA4E;AAC5E,2EAA2E;AAC3E,2EAA2E;AAC3E,wEAAwE;AACxE,wEAAwE;AACxE,+EAA+E;AAC/E,EAAE;AACF,kDAAkD;AAClD,EAAE;AACF,sBAAsB;AACtB,wEAAwE;AACxE,yEAAyE;AACzE,oEAAoE;AACpE,wEAAwE;AACxE,uEAAuE;AACvE,2EAA2E;AAC3E,wEAAwE;AACxE,0EAA0E;AAC1E,EAAE;AACF,yEAAyE;AACzE,uEAAuE;AACvE,uEAAuE;AACvE,sEAAsE;AACtE,wEAAwE;AACxE,4DAA4D;AAC5D,EAAE;AACF,mEAAmE;AACnE,sEAAsE;AACtE,qEAAqE;AACrE,kEAAkE;AAClE,EAAE;AACF,iEAAiE;AACjE,wEAAwE;AACxE,oEAAoE;AACpE,EAAE;AACF,sEAAsE;AACtE,0EAA0E;AAC1E,iEAAiE;AACjE,WAAW;AAuJX,IAAI,YAAY,GAA6B,IAAI,CAAC;AAClD,KAAK,UAAU,WAAW;IACxB,IAAI,YAAY;QAAE,OAAO,YAAY,CAAC;IACtC,IAAI,CAAC;QACH,MAAM,GAAG,GAAG,CAAC,MAAM,MAAM,CAAC,cAAc,CAAC,CAAiE,CAAC;QAC3G,qEAAqE;QACrE,mDAAmD;QACnD,MAAM,GAAG,GAAG,GAAG,CAAC,OAAO,IAAK,GAAyB,CAAC;QACtD,IAAI,OAAO,GAAG,CAAC,eAAe,KAAK,UAAU,EAAE,CAAC;YAC9C,MAAM,IAAI,KAAK,CAAC,+DAA+D,CAAC,CAAC;QACnF,CAAC;QACD,YAAY,GAAG,GAAG,CAAC;QACnB,OAAO,YAAY,CAAC;IACtB,CAAC;IAAC,OAAO,GAAG,EAAE,CAAC;QACb,MAAM,GAAG,GAAG,GAAG,YAAY,KAAK,CAAC,CAAC,CAAC,GAAG,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,GAAG,CAAC,CAAC;QAC7D,MAAM,IAAI,KAAK,CACb,kFAAkF;YAChF,gGAAgG;YAChG,qBAAqB,GAAG,EAAE,CAC7B,CAAC;IACJ,CAAC;AACH,CAAC;AAED;;;;;;;;;;GAUG;AACH,MAAM,CAAC,KAAK,UAAU,SAAS,CAAC,OAAqC,EAAE,IAAsB;IAC3F,MAAM,GAAG,GAAG,IAAI,CAAC,GAAG,CAAC;IACrB,IAAI,OAAO,CAAC,MAAM,GAAG,IAAI,CAAC,WAAW,EAAE,CAAC;QACtC,MAAM,IAAI,KAAK,CACb,6BAA6B,OAAO,CAAC,MAAM,wBAAwB,IAAI,CAAC,WAAW,sBAAsB,CAC1G,CAAC;IACJ,CAAC;IACD,MAAM,CAAC,GAAG,IAAI,CAAC,CAAC,IAAI,EAAE,CAAC;IACvB,MAAM,cAAc,GAAG,IAAI,CAAC,cAAc,IAAI,GAAG,CAAC;IAClD,MAAM,IAAI,GAAG,IAAI,CAAC,IAAI,IAAI,GAAG,CAAC;IAE9B,gEAAgE;IAChE,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,OAAO,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;QACxC,MAAM,CAAC,GAAG,OAAO,CAAC,CAAC,CAAC,CAAC;QACrB,IAAI,CAAC,CAAC;YAAE,SAAS;QACjB,IAAI,CAAC,CAAC,MAAM,CAAC,MAAM,KAAK,GAAG,EAAE,CAAC;YAC5B,MAAM,IAAI,KAAK,CAAC,8BAA8B,CAAC,YAAY,CAAC,CAAC,MAAM,CAAC,MAAM,cAAc,GAAG,EAAE,CAAC,CAAC;QACjG,CAAC;IACH,CAAC;IAED,MAAM,GAAG,GAAG,MAAM,WAAW,EAAE,CAAC;IAChC,MAAM,IAAI,GAAG,IAAI,GAAG,CAAC,eAAe,CAAC,QAAQ,EAAE,GAAG,CAAC,CAAC;IACpD,wEAAwE;IACxE,yEAAyE;IACzE,uBAAuB;IACvB,IAAI,CAAC,SAAS,CAAC,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,WAAW,EAAE,CAAC,CAAC,EAAE,CAAC,EAAE,cAAc,EAAE,IAAI,CAAC,CAAC;IAEvE,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,OAAO,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;QACxC,MAAM,CAAC,GAAG,OAAO,CAAC,CAAC,CAAC,CAAC;QACrB,IAAI,CAAC,CAAC;YAAE,SAAS;QACjB,kEAAkE;QAClE,sEAAsE;QACtE,qEAAqE;QACrE,IAAI,CAAC,QAAQ,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,CAAC,MAAM,CAAC,EAAE,CAAC,CAAC,KAAK,CAAC,CAAC;IAC/C,CAAC;IAED,OAAO,eAAe,CAAC,IAAI,EAAE,GAAG,EAAE,OAAO,CAAC,MAAM,CAAC,CAAC;AACpD,CAAC;AAED;;;;GAIG;AACH,SAAS,eAAe,CAAC,IAAqB,EAAE,GAAW,EAAE,IAAY;IACvE,OAAO;QACL,GAAG;QACH,IAAI;QACJ,SAAS,CAAC,QAAsB,EAAE,CAAS,EAAE,KAAwB;YACnE,IAAI,QAAQ,CAAC,MAAM,KAAK,GAAG,EAAE,CAAC;gBAC5B,MAAM,IAAI,KAAK,CAAC,kCAAkC,QAAQ,CAAC,MAAM,gBAAgB,GAAG,EAAE,CAAC,CAAC;YAC1F,CAAC;YACD,oEAAoE;YACpE,0CAA0C;YAC1C,MAAM,EAAE,GAAG,IAAI,CAAC,GAAG,CAAC,KAAK,EAAE,EAAE,IAAI,GAAG,EAAE,CAAC,CAAC,CAAC;YACzC,IAAI,CAAC,KAAK,CAAC,EAAE,CAAC,CAAC;YACf,MAAM,MAAM,GAAG,IAAI,CAAC,SAAS,CAAC,KAAK,CAAC,IAAI,CAAC,QAAQ,CAAC,EAAE,CAAC,EAAE,SAAS,CAAC,CAAC;YAClE,OAAO,EAAE,MAAM,EAAE,MAAM,CAAC,SAAS,EAAE,SAAS,EAAE,MAAM,CAAC,SAAS,EAAE,CAAC;QACnE,CAAC;QACD,KAAK,CAAC,MAAM,CAAC,IAAI,EAAE,WAAW,EAAE,SAAS;YACvC,MAAM,EAAE,GAAG,MAAM,MAAM,CAAC,kBAAkB,CAAC,CAAC;YAC5C,MAAM,IAAI,GAAG,MAAM,MAAM,CAAC,WAAW,CAAC,CAAC;YACvC,iEAAiE;YACjE,iEAAiE;YACjE,0DAA0D;YAC1D,MAAM,EAAE,CAAC,KAAK,CAAC,IAAI,CAAC,OAAO,CAAC,IAAI,CAAC,EAAE,EAAE,SAAS,EAAE,IAAI,EAAE,CAAC,CAAC;YACxD,MAAM,OAAO,GAAG,GAAG,IAAI,MAAM,CAAC;YAC9B,MAAM,QAAQ,GAAG,GAAG,IAAI,YAAY,CAAC;YACrC,6DAA6D;YAC7D,gEAAgE;YAChE,MAAM,IAAI,CAAC,UAAU,CAAC,OAAO,CAAC,CAAC;YAC/B,MAAM,IAAI,GAAsB;gBAC9B,aAAa,EAAE,CAAC;gBAChB,GAAG;gBACH,IAAI;gBACJ,SAAS;gBACT,WAAW,EAAE,MAAM,CAAC,WAAW,CAAC,WAAW,CAAC;gBAC5C,SAAS,EAAE,IAAI,IAAI,EAAE,CAAC,WAAW,EAAE;aACpC,CAAC;YACF,MAAM,EAAE,CAAC,SAAS,CAAC,QAAQ,EAAE,IAAI,CAAC,SAAS,CAAC,IAAI,EAAE,IAAI,EAAE,CAAC,CAAC,EAAE,MAAM,CAAC,CAAC;YACpE,OAAO,IAAI,CAAC;QACd,CAAC;KACF,CAAC;AACJ,CAAC;AAED;;;;;;;;;;;;GAYG;AACH,MAAM,CAAC,KAAK,UAAU,gBAAgB,CACpC,IAAY,EACZ,iBAAyB;IAEzB,MAAM,EAAE,GAAG,MAAM,MAAM,CAAC,kBAAkB,CAAC,CAAC;IAC5C,MAAM,OAAO,GAAG,GAAG,IAAI,MAAM,CAAC;IAC9B,MAAM,QAAQ,GAAG,GAAG,IAAI,YAAY,CAAC;IACrC,IAAI,OAAe,CAAC;IACpB,IAAI,CAAC;QACH,OAAO,GAAG,MAAM,EAAE,CAAC,QAAQ,CAAC,QAAQ,EAAE,MAAM,CAAC,CAAC;IAChD,CAAC;IAAC,MAAM,CAAC;QACP,OAAO,IAAI,CAAC,CAAC,mDAAmD;IAClE,CAAC;IACD,IAAI,IAAuB,CAAC;IAC5B,IAAI,CAAC;QACH,IAAI,GAAG,IAAI,CAAC,KAAK,CAAC,OAAO,CAAsB,CAAC;IAClD,CAAC;IAAC,OAAO,GAAG,EAAE,CAAC;QACb,OAAO,CAAC,MAAM,CAAC,KAAK,CAClB,yBAAyB,QAAQ,+BAA+B,GAAG,YAAY,KAAK,CAAC,CAAC,CAAC,GAAG,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,GAAG,CAAC,IAAI,CACrH,CAAC;QACF,OAAO,IAAI,CAAC;IACd,CAAC;IACD,IAAI,IAAI,CAAC,aAAa,KAAK,CAAC,EAAE,CAAC;QAC7B,OAAO,CAAC,MAAM,CAAC,KAAK,CAClB,6BAA6B,IAAI,CAAC,aAAa,mEAAmE,CACnH,CAAC;QACF,OAAO,IAAI,CAAC;IACd,CAAC;IACD,IAAI,IAAI,CAAC,SAAS,KAAK,iBAAiB,EAAE,CAAC;QACzC,OAAO,CAAC,MAAM,CAAC,KAAK,CAClB,+GAA+G,CAChH,CAAC;QACF,OAAO,IAAI,CAAC;IACd,CAAC;IACD,oBAAoB;IACpB,IAAI,CAAC;QACH,MAAM,EAAE,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC;IAC3B,CAAC;IAAC,MAAM,CAAC;QACP,OAAO,CAAC,MAAM,CAAC,KAAK,CAAC,iCAAiC,OAAO,2BAA2B,CAAC,CAAC;QAC1F,OAAO,IAAI,CAAC;IACd,CAAC;IACD,0BAA0B;IAC1B,MAAM,GAAG,GAAG,MAAM,WAAW,EAAE,CAAC;IAChC,MAAM,IAAI,GAAG,IAAI,GAAG,CAAC,eAAe,CAAC,QAAQ,EAAE,IAAI,CAAC,GAAG,CAAC,CAAC;IACzD,IAAI,CAAC;QACH,MAAM,IAAI,CAAC,SAAS,CAAC,OAAO,CAAC,CAAC;IAChC,CAAC;IAAC,OAAO,GAAG,EAAE,CAAC;QACb,OAAO,CAAC,MAAM,CAAC,KAAK,CAClB,qCAAqC,OAAO,kBAAkB,GAAG,YAAY,KAAK,CAAC,CAAC,CAAC,GAAG,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,GAAG,CAAC,IAAI,CACnH,CAAC;QACF,OAAO,IAAI,CAAC;IACd,CAAC;IACD,MAAM,KAAK,GAAG,eAAe,CAAC,IAAI,EAAE,IAAI,CAAC,GAAG,EAAE,IAAI,CAAC,IAAI,CAAC,CAAC;IACzD,2BAA2B;IAC3B,MAAM,WAAW,GAAG,IAAI,GAAG,EAAoD,CAAC;IAChF,KAAK,MAAM,CAAC,QAAQ,EAAE,GAAG,CAAC,IAAI,MAAM,CAAC,OAAO,CAAC,IAAI,CAAC,WAAW,CAAC,EAAE,CAAC;QAC/D,WAAW,CAAC,GAAG,CAAC,MAAM,CAAC,QAAQ,CAAC,QAAQ,EAAE,EAAE,CAAC,EAAE,GAAG,CAAC,CAAC;IACtD,CAAC;IACD,OAAO,EAAE,KAAK,EAAE,WAAW,EAAE,CAAC;AAChC,CAAC;AAED;;;;;;;GAOG;AACH,MAAM,UAAU,iBAAiB,CAC/B,MAAiD,EACjD,UAUC;IAED,MAAM,IAAI,GAAqB,EAAE,CAAC;IAClC,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,MAAM,CAAC,MAAM,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;QAC9C,MAAM,KAAK,GAAG,MAAM,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC;QAC/B,MAAM,QAAQ,GAAG,MAAM,CAAC,SAAS,CAAC,CAAC,CAAC,CAAC;QACrC,IAAI,KAAK,KAAK,SAAS,IAAI,QAAQ,KAAK,SAAS;YAAE,SAAS;QAC5D,MAAM,GAAG,GAAG,UAAU,CAAC,GAAG,CAAC,KAAK,CAAC,CAAC;QAClC,IAAI,CAAC,GAAG;YAAE,SAAS,CAAC,mDAAmD;QACvE,wDAAwD;QACxD,kEAAkE;QAClE,MAAM,KAAK,GAAG,CAAC,GAAG,QAAQ,CAAC;QAC3B,IAAI,CAAC,IAAI,CAAC;YACR,QAAQ,EAAE,GAAG,CAAC,QAAQ;YACtB,WAAW,EAAE,GAAG,CAAC,WAAW;YAC5B,UAAU,EAAE,GAAG,CAAC,UAAU;YAC1B,QAAQ,EAAE,GAAG,CAAC,QAAQ;YACtB,YAAY,EAAE,GAAG,CAAC,YAAY;YAC9B,KAAK;YACL,IAAI,EAAE,GAAG,CAAC,IAAI;SACf,CAAC,CAAC;IACL,CAAC;IACD,OAAO,IAAI,CAAC;AACd,CAAC"}

package/dist/index.d.ts

@@ -39,6 +39,19 @@ export interface ServeOptions {
     hnswEf?: string;
     /** v2.15.0 — late-chunking context windowing for embeddings (default 0 chars). */
     lateChunkContext?: string;
+    /** v2.16.0 — persist HNSW index to disk for fast reload on next serve.
+     *  Default true (the persistence is a pure optimization; corrupt files
+     *  fall back to rebuild gracefully). Pass `--no-hnsw-persist` to opt out. */
+    hnswPersist?: boolean;
+    /** v2.17.0 — vector storage encoding for the persistent embed db.
+     *  - `"f32"` (default) — Float32 BLOB, identical to v2.16- behavior.
+     *  - `"int8"` — int8-quantized BLOB + per-vector (vMin, scale) Float32
+     *    tuple. ~4× storage reduction at ~1-2% recall@10 cost.
+     *  Mode is per-database; switching modes triggers a full rebuild
+     *  (the meta-table contamination guard treats it as a schema change).
+     *  Must match the mode used at build-embeddings time — serving with a
+     *  different mode would auto-rebuild the index. */
+    quantizeEmbeddings?: "f32" | "int8";
 }
 declare function main(): Promise<void>;
 /**
@@ -128,5 +141,12 @@ export declare function buildEmbedText(chunks: ReadonlyArray<{
     contextChars: number;
 }): string;
 declare function parsePositiveInt(raw: string, flag: string): number;
-export { main, parsePositiveInt, startServer };
+/**
+ * v2.17.0 — validate a `--quantize-embeddings <mode>` value. Accepts the
+ * canonical `"f32"` / `"int8"` plus a few user-friendly aliases (`"none"`
+ * for f32; `"q8"` for int8). Anything else throws with the exact list of
+ * accepted values so the user can fix the typo immediately.
+ */
+declare function parseQuantizationMode(raw: string | undefined): "f32" | "int8" | undefined;
+export { main, parsePositiveInt, parseQuantizationMode, startServer };
 //# sourceMappingURL=index.d.ts.map

package/dist/index.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../src/index.ts"],"names":[],"mappings":";AAIA,OAAO,EAAE,SAAS,EAAoB,MAAM,yCAAyC,CAAC;AAMtF,OAAO,EAAkC,QAAQ,EAAE,MAAM,WAAW,CAAC;AAyCrE,OAAO,EAAE,KAAK,EAAE,MAAM,YAAY,CAAC;AACnC,OAAO,EAAE,YAAY,EAAE,MAAM,cAAc,CAAC;AAW5C,MAAM,WAAW,YAAY;IAC3B,KAAK,EAAE,MAAM,CAAC;IACd,WAAW,CAAC,EAAE,OAAO,CAAC;IACtB,YAAY,CAAC,EAAE,MAAM,CAAC;IACtB,SAAS,CAAC,EAAE,MAAM,CAAC;IACnB,eAAe,CAAC,EAAE,OAAO,CAAC;IAC1B,SAAS,CAAC,EAAE,MAAM,CAAC;IACnB,eAAe,CAAC,EAAE,OAAO,CAAC;IAC1B,SAAS,CAAC,EAAE,MAAM,CAAC;IACnB,QAAQ,CAAC,EAAE,WAAW,GAAG,SAAS,CAAC;IACnC,WAAW,CAAC,EAAE,MAAM,EAAE,CAAC;IACvB,SAAS,CAAC,EAAE,MAAM,EAAE,CAAC;IACrB,KAAK,CAAC,EAAE,OAAO,CAAC;IAChB,aAAa,CAAC,EAAE,MAAM,EAAE,CAAC;IACzB,YAAY,CAAC,EAAE,MAAM,EAAE,CAAC;IACxB,qBAAqB,CAAC,EAAE,OAAO,CAAC;IAChC;;mCAE+B;IAC/B,WAAW,CAAC,EAAE,OAAO,CAAC;IACtB;8EAC0E;IAC1E,cAAc,CAAC,EAAE,OAAO,CAAC;IACzB,qEAAqE;IACrE,aAAa,CAAC,EAAE,MAAM,CAAC;IACvB,qEAAqE;IACrE,YAAY,CAAC,EAAE,MAAM,CAAC;IACtB;;;+BAG2B;IAC3B,OAAO,CAAC,EAAE,OAAO,CAAC;IAClB,+DAA+D;IAC/D,MAAM,CAAC,EAAE,MAAM,CAAC;IAChB,kFAAkF;IAClF,gBAAgB,CAAC,EAAE,MAAM,CAAC;CAC3B;AAkBD,iBAAe,IAAI,IAAI,OAAO,CAAC,IAAI,CAAC,CAsnBnC;AAED;;;;;;;;;;GAUG;AACH,MAAM,WAAW,UAAU;IACzB,KAAK,EAAE,KAAK,CAAC;IACb,QAAQ,EAAE,QAAQ,GAAG,IAAI,CAAC;IAC1B,OAAO,EAAE,YAAY,GAAG,IAAI,CAAC;IAC7B,aAAa,EAAE,GAAG,CAAC,MAAM,CAAC,CAAC;IAC3B,YAAY,EAAE,GAAG,CAAC,MAAM,CAAC,CAAC;IAC1B,cAAc,EAAE;QAAE,OAAO,EAAE,OAAO,CAAA;KAAE,CAAC;IACrC;;;;OAIG;IACH,WAAW,EAAE;QACX,sBAAsB;QACtB,KAAK,EAAE,OAAO,WAAW,EAAE,SAAS,CAAC;QACrC,oEAAoE;QACpE,UAAU,EAAE,GAAG,CACb,MAAM,EACN;YACE,QAAQ,EAAE,MAAM,CAAC;YACjB,WAAW,EAAE,MAAM,CAAC;YACpB,UAAU,EAAE,MAAM,CAAC;YACnB,QAAQ,EAAE,MAAM,CAAC;YACjB,YAAY,EAAE,MAAM,CAAC;YACrB,IAAI,EAAE,IAAI,GAAG,KAAK,CAAC;SACpB,CACF,CAAC;QACF,kFAAkF;QAClF,EAAE,CAAC,EAAE,MAAM,CAAC;KACb,GAAG,IAAI,CAAC;CACV;AAED;;;;;GAKG;AACH,wBAAsB,iBAAiB,CAAC,IAAI,EAAE,YAAY,GAAG,OAAO,CAAC,UAAU,CAAC,CAgJ/E;AAED;;;;GAIG;AACH,wBAAgB,cAAc,CAAC,IAAI,EAAE,UAAU,EAAE,IAAI,EAAE,YAAY,GAAG,SAAS,CAqG9E;AAED,iBAAe,WAAW,CAAC,IAAI,EAAE,YAAY,GAAG,OAAO,CAAC,IAAI,CAAC,CAoD5D;AAED;;;;GAIG;AACH,wBAAgB,iBAAiB,CAAC,IAAI,EAAE,UAAU,GAAG,MAAM,CAY1D;AAED;;;;;;;;;;;;;;;;;GAiBG;AACH,wBAAgB,cAAc,CAC5B,MAAM,EAAE,aAAa,CAAC;IAAE,IAAI,EAAE,MAAM,CAAC;IAAC,UAAU,CAAC,EAAE,MAAM,CAAA;CAAE,CAAC,EAC5D,CAAC,EAAE,MAAM,EACT,IAAI,EAAE;IAAE,QAAQ,CAAC,EAAE,MAAM,CAAC;IAAC,YAAY,EAAE,MAAM,CAAA;CAAE,GAChD,MAAM,CAwBR;AAi/DD,iBAAS,gBAAgB,CAAC,GAAG,EAAE,MAAM,EAAE,IAAI,EAAE,MAAM,GAAG,MAAM,CAM3D;AAsCD,OAAO,EAAE,IAAI,EAAE,gBAAgB,EAAE,WAAW,EAAE,CAAC"}
+{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../src/index.ts"],"names":[],"mappings":";AAIA,OAAO,EAAE,SAAS,EAAoB,MAAM,yCAAyC,CAAC;AAMtF,OAAO,EAAkC,QAAQ,EAAE,MAAM,WAAW,CAAC;AAyCrE,OAAO,EAAE,KAAK,EAAE,MAAM,YAAY,CAAC;AACnC,OAAO,EAAE,YAAY,EAAE,MAAM,cAAc,CAAC;AAW5C,MAAM,WAAW,YAAY;IAC3B,KAAK,EAAE,MAAM,CAAC;IACd,WAAW,CAAC,EAAE,OAAO,CAAC;IACtB,YAAY,CAAC,EAAE,MAAM,CAAC;IACtB,SAAS,CAAC,EAAE,MAAM,CAAC;IACnB,eAAe,CAAC,EAAE,OAAO,CAAC;IAC1B,SAAS,CAAC,EAAE,MAAM,CAAC;IACnB,eAAe,CAAC,EAAE,OAAO,CAAC;IAC1B,SAAS,CAAC,EAAE,MAAM,CAAC;IACnB,QAAQ,CAAC,EAAE,WAAW,GAAG,SAAS,CAAC;IACnC,WAAW,CAAC,EAAE,MAAM,EAAE,CAAC;IACvB,SAAS,CAAC,EAAE,MAAM,EAAE,CAAC;IACrB,KAAK,CAAC,EAAE,OAAO,CAAC;IAChB,aAAa,CAAC,EAAE,MAAM,EAAE,CAAC;IACzB,YAAY,CAAC,EAAE,MAAM,EAAE,CAAC;IACxB,qBAAqB,CAAC,EAAE,OAAO,CAAC;IAChC;;mCAE+B;IAC/B,WAAW,CAAC,EAAE,OAAO,CAAC;IACtB;8EAC0E;IAC1E,cAAc,CAAC,EAAE,OAAO,CAAC;IACzB,qEAAqE;IACrE,aAAa,CAAC,EAAE,MAAM,CAAC;IACvB,qEAAqE;IACrE,YAAY,CAAC,EAAE,MAAM,CAAC;IACtB;;;+BAG2B;IAC3B,OAAO,CAAC,EAAE,OAAO,CAAC;IAClB,+DAA+D;IAC/D,MAAM,CAAC,EAAE,MAAM,CAAC;IAChB,kFAAkF;IAClF,gBAAgB,CAAC,EAAE,MAAM,CAAC;IAC1B;;iFAE6E;IAC7E,WAAW,CAAC,EAAE,OAAO,CAAC;IACtB;;;;;;;uDAOmD;IACnD,kBAAkB,CAAC,EAAE,KAAK,GAAG,MAAM,CAAC;CACrC;AAkBD,iBAAe,IAAI,IAAI,OAAO,CAAC,IAAI,CAAC,CAqqBnC;AAED;;;;;;;;;;GAUG;AACH,MAAM,WAAW,UAAU;IACzB,KAAK,EAAE,KAAK,CAAC;IACb,QAAQ,EAAE,QAAQ,GAAG,IAAI,CAAC;IAC1B,OAAO,EAAE,YAAY,GAAG,IAAI,CAAC;IAC7B,aAAa,EAAE,GAAG,CAAC,MAAM,CAAC,CAAC;IAC3B,YAAY,EAAE,GAAG,CAAC,MAAM,CAAC,CAAC;IAC1B,cAAc,EAAE;QAAE,OAAO,EAAE,OAAO,CAAA;KAAE,CAAC;IACrC;;;;OAIG;IACH,WAAW,EAAE;QACX,sBAAsB;QACtB,KAAK,EAAE,OAAO,WAAW,EAAE,SAAS,CAAC;QACrC,oEAAoE;QACpE,UAAU,EAAE,GAAG,CACb,MAAM,EACN;YACE,QAAQ,EAAE,MAAM,CAAC;YACjB,WAAW,EAAE,MAAM,CAAC;YACpB,UAAU,EAAE,MAAM,CAAC;YACnB,QAAQ,EAAE,MAAM,CAAC;YACjB,YAAY,EAAE,MAAM,CAAC;YACrB,IAAI,EAAE,IAAI,GAAG,KAAK,CAAC;SACpB,CACF,CAAC;QACF,kFAAkF;QAClF,EAAE,CAAC,EAAE,MAAM,CAAC;KACb,GAAG,IAAI,CAAC;CACV;AAED;;;;;GAKG;AACH,wBAAsB,iBAAiB,CAAC,IAAI,EAAE,YAAY,GAAG,OAAO,CAAC,UAAU,CAAC,CA4M/E;AAED;;;;GAIG;AACH,wBAAgB,cAAc,CAAC,IAAI,EAAE,UAAU,EAAE,IAAI,EAAE,YAAY,GAAG,SAAS,CAqG9E;AAED,iBAAe,WAAW,CAAC,IAAI,EAAE,YAAY,GAAG,OAAO,CAAC,IAAI,CAAC,CAoD5D;AAED;;;;GAIG;AACH,wBAAgB,iBAAiB,CAAC,IAAI,EAAE,UAAU,GAAG,MAAM,CAY1D;AAED;;;;;;;;;;;;;;;;;GAiBG;AACH,wBAAgB,cAAc,CAC5B,MAAM,EAAE,aAAa,CAAC;IAAE,IAAI,EAAE,MAAM,CAAC;IAAC,UAAU,CAAC,EAAE,MAAM,CAAA;CAAE,CAAC,EAC5D,CAAC,EAAE,MAAM,EACT,IAAI,EAAE;IAAE,QAAQ,CAAC,EAAE,MAAM,CAAC;IAAC,YAAY,EAAE,MAAM,CAAA;CAAE,GAChD,MAAM,CAwBR;AAi/DD,iBAAS,gBAAgB,CAAC,GAAG,EAAE,MAAM,EAAE,IAAI,EAAE,MAAM,GAAG,MAAM,CAM3D;AAED;;;;;GAKG;AACH,iBAAS,qBAAqB,CAAC,GAAG,EAAE,MAAM,GAAG,SAAS,GAAG,KAAK,GAAG,MAAM,GAAG,SAAS,CASlF;AAsCD,OAAO,EAAE,IAAI,EAAE,gBAAgB,EAAE,qBAAqB,EAAE,WAAW,EAAE,CAAC"}

package/dist/index.js

@@ -12,7 +12,7 @@ import { chunkContent, defaultIndexFile, FtsIndex } from "./fts5.js";
 import { appendToNote, archiveNote, chatThreadAppend, chatThreadRead, contextPack, createNote, dataviewQuery, embeddingsSearch, findPath, findSimilar, frontmatterGet, frontmatterSearch, frontmatterSet, getBacklinks, getNoteNeighbors, getOpenQuestions, getOutboundLinks, getRecentEdits, getUnresolvedWikilinks, getVaultStats, lintWiki, listCanvases, listNotes, listPdfs, listTags, ocrPdf, openInUi, paperAudit, readCanvas, readNote, readPdf, renameNote, replaceInNotes, resolveWikilink, searchHybrid, searchText, semanticSearch, validateNoteProposal } from "./tools.js";
 import { Vault } from "./vault.js";
 import { VaultWatcher } from "./watcher.js";
-const VERSION = "2.15.0";
+const VERSION = "2.17.0";
 /** Default location for the persistent embedding index, alongside .fts5.db. */
 function embedDbPath(vaultRoot) {
     // Match the FTS5 location convention by stripping the .fts5.db extension
@@ -50,7 +50,11 @@ async function main() {
         .option("--use-hnsw", "v2.13.0 — build an in-memory HNSW vector index on serve start (or rebuild if `.embed.db` is missing). Sub-10ms top-K queries at any vault scale, vs O(n) brute-force without it. Build cost: ~5s for 8K chunks, ~25s for 50K, ~4min for 500K (one-time per serve). Recall@10 ≥ 98% vs brute-force at default params. Requires the `hnswlib-wasm` optionalDependency (~340 KB, pure WASM, no native binding).")
         .option("--hnsw-ef <n>", "v2.13.0 — HNSW search-time beam width (default 100; must be ≥ requested k). Higher = more accurate, slightly slower. Common range: 50-500. Only effective with `--use-hnsw`.")
         .option("--late-chunk-context <chars>", "v2.15.0 — late-chunking-style context windowing on embeddings. When > 0, prepends doc title + heading breadcrumb + tails of neighboring chunks (this many chars from each side) before sending to the embedder. Typical +2-5 NDCG@10 retrieval boost at zero new dep cost. Default 0 (off; matches v2.1.0+ breadcrumb-only behavior). Only effective during `build-embeddings` or auto-rebuild.")
+        .option("--no-hnsw-persist", "v2.16.0 — disable HNSW index persistence. By default (with --use-hnsw), the index is saved to a sidecar `.hnsw.bin` + `.meta.json` next to `.embed.db` after the first build, then re-loaded on subsequent serve starts when the embed-db signature matches. Skipping persistence means a fresh rebuild every serve start (~25s for 50K chunks). Pass this flag if you can't write to the cache dir or want diagnostic-fresh builds.")
+        .option("--quantize-embeddings <mode>", "v2.17.0 — vector storage encoding for the persistent embed db. `f32` (default) is identical to v2.16- behavior. `int8` cuts BLOB size ~4× (per-vector min+scale + int8 bytes) at ~1-2% recall@10 cost. Must match the mode used at `build-embeddings` time — otherwise the index auto-rebuilds on serve start. Accepts `f32`/`float32`/`none` and `int8`/`i8`/`q8`.")
         .action(async (opts) => {
+        // Validate up-front so a bad value fails before we touch the vault.
+        parseQuantizationMode(opts.quantizeEmbeddings);
         await startServer(opts);
     });
     // v2.6.0 — remote-MCP HTTP transport. Mirrors `serve` flags + adds HTTP
@@ -84,6 +88,7 @@ async function main() {
         .option("--disabled-tools <name...>", "Skip registration of specific tools by name.")
         .option("--enabled-tools <name...>", "Strict allowlist — when set, ONLY listed tools register.")
         .option("--diagnostic-search-tools", "Register the four single-ranker search tools alongside obsidian_search.")
+        .option("--quantize-embeddings <mode>", "v2.17.0 — vector storage encoding for the persistent embed db (`f32` default, `int8` for ~4× smaller BLOBs). Must match the mode used at `build-embeddings` time.")
         .action(async (opts) => {
         const tokenFromArg = typeof opts.bearerToken === "string" ? opts.bearerToken.trim() : "";
         const tokenFromEnv = typeof opts.bearerTokenEnv === "string" ? (process.env[opts.bearerTokenEnv] ?? "").trim() : "";
@@ -107,8 +112,11 @@ async function main() {
             ? parsePositiveInt(opts.sessionIdleTimeoutMs, "--session-idle-timeout-ms")
             : 30 * 60 * 1000;
         const maxSessionsCap = opts.maxSessions !== undefined ? parsePositiveInt(opts.maxSessions, "--max-sessions") : 100;
+        // v2.17.0 — fail fast on a typo'd quantization mode.
+        const quantMode = parseQuantizationMode(opts.quantizeEmbeddings);
         const httpOpts = {
             ...opts,
+            ...(quantMode !== undefined ? { quantizeEmbeddings: quantMode } : {}),
             port: portNum,
             host: opts.host ?? "127.0.0.1",
             bearerToken,
@@ -229,12 +237,20 @@ async function main() {
         .option("--read-paths <pattern...>", "Strict allowlist of glob patterns (repeatable)")
         .option("--include-pdfs", "v2.8.0 — also embed PDF chunks. Off by default; PDF extraction + embedding is ~10-30x slower than markdown per file.")
         .option("--late-chunk-context <chars>", "v2.15.0 — context-windowed embedding text (doc title + breadcrumb + neighbor-chunk tails of N chars). Default 0 (off). Typical 100-200 for +2-5 NDCG@10.")
+        .option("--quantize-embeddings <mode>", "v2.17.0 — vector storage encoding. `f32` (default) is identical to v2.16- behavior. `int8` uses asymmetric scalar quantization (per-vector min + scale + int8 bytes) for ~4× smaller BLOBs at ~1-2% recall@10 cost. Switching modes triggers a full rebuild via the schema-mismatch path. Accepts `f32`/`float32`/`none` and `int8`/`i8`/`q8`.")
         .action(async (opts) => {
         const model = resolveModel(opts.embeddingModel);
         const vault = new Vault(opts.vault, { excludeGlobs: opts.excludeGlob, readPaths: opts.readPaths });
         await vault.ensureExists();
         const embedFile = opts.embedFile ?? embedDbPath(vault.root);
-        const db = new EmbedDb({ file: embedFile, vaultRoot: vault.root, modelAlias: model.alias, dim: model.dim });
+        const quantization = parseQuantizationMode(opts.quantizeEmbeddings) ?? "f32";
+        const db = new EmbedDb({
+            file: embedFile,
+            vaultRoot: vault.root,
+            modelAlias: model.alias,
+            dim: model.dim,
+            quantization
+        });
         const lateChunkContext = opts.lateChunkContext !== undefined
             ? Math.max(0, parsePositiveInt(opts.lateChunkContext, "--late-chunk-context"))
             : 0;
@@ -243,7 +259,7 @@ async function main() {
             process.stderr.write(`enquire: loading embedder ${model.alias} (${model.hfId})...\n`);
             const embedder = await loadEmbedder(opts.embeddingModel);
             const report = await syncEmbedDb(vault, db, embedder, { lateChunkContext });
-            process.stdout.write(`enquire: embed db ${embedFile} (md) — added=${report.added} updated=${report.updated} deleted=${report.deleted} unchanged=${report.unchanged} total_chunks=${report.total_chunks}${lateChunkContext > 0 ? ` late-chunk-context=${lateChunkContext}` : ""}\n`);
+            process.stdout.write(`enquire: embed db ${embedFile} (md) — added=${report.added} updated=${report.updated} deleted=${report.deleted} unchanged=${report.unchanged} total_chunks=${report.total_chunks}${lateChunkContext > 0 ? ` late-chunk-context=${lateChunkContext}` : ""}${quantization !== "f32" ? ` quantization=${quantization}` : ""}\n`);
             if (opts.includePdfs) {
                 const pdfReport = await syncPdfEmbedDb(vault, db, embedder, { lateChunkContext });
                 process.stdout.write(`enquire: embed db ${embedFile} (pdf) — added=${pdfReport.added} updated=${pdfReport.updated} deleted=${pdfReport.deleted} unchanged=${pdfReport.unchanged} total_chunks=${pdfReport.total_chunks}\n`);
@@ -303,6 +319,7 @@ async function main() {
         .option("--embedding-model <alias>", `Model alias (default: ${DEFAULT_MODEL_ALIAS})`, DEFAULT_MODEL_ALIAS)
         .option("--include-pdfs", "Also index PDFs (FTS5 + embeddings). Off by default; opt-in because PDF extraction is slower.")
         .option("--skip-embeddings", "Skip the install-model + build-embeddings steps (only build FTS5)")
+        .option("--quantize-embeddings <mode>", "v2.17.0 — vector storage encoding for the embed db (`f32` default, `int8` for ~4× smaller BLOBs). Same semantics as the `build-embeddings` flag.")
         .action(async (opts) => {
         const v = new Vault(opts.vault);
         await v.ensureExists();
@@ -341,11 +358,18 @@ async function main() {
         // Step 3: build-embeddings.
         process.stdout.write("\n>> Step 3/3: Build embedding index\n");
         const embedFile = embedDbPath(v.root);
-        const db = new EmbedDb({ file: embedFile, vaultRoot: v.root, modelAlias: model.alias, dim: model.dim });
+        const quantization = parseQuantizationMode(opts.quantizeEmbeddings) ?? "f32";
+        const db = new EmbedDb({
+            file: embedFile,
+            vaultRoot: v.root,
+            modelAlias: model.alias,
+            dim: model.dim,
+            quantization
+        });
         await db.open();
         try {
             const embReport = await syncEmbedDb(v, db, embedder);
-            process.stdout.write(`   embed-db (md): added=${embReport.added} updated=${embReport.updated} unchanged=${embReport.unchanged} chunks=${embReport.total_chunks}\n`);
+            process.stdout.write(`   embed-db (md): added=${embReport.added} updated=${embReport.updated} unchanged=${embReport.unchanged} chunks=${embReport.total_chunks}${quantization !== "f32" ? ` quantization=${quantization}` : ""}\n`);
             if (opts.includePdfs) {
                 const pdfReport = await syncPdfEmbedDb(v, db, embedder);
                 process.stdout.write(`   embed-db (pdf): added=${pdfReport.added} updated=${pdfReport.updated} unchanged=${pdfReport.unchanged} chunks=${pdfReport.total_chunks}\n`);
@@ -358,6 +382,8 @@ async function main() {
         process.stdout.write(`   enquire-mcp serve --vault ${opts.vault} --persistent-index`);
         if (opts.includePdfs)
             process.stdout.write(" --include-pdfs");
+        if (quantization !== "f32")
+            process.stdout.write(` --quantize-embeddings ${quantization}`);
         process.stdout.write("\n");
         process.stdout.write(`Or check status: enquire-mcp doctor --vault ${opts.vault}\n`);
     });
@@ -562,31 +588,76 @@ export async function prepareServerDeps(opts) {
                 // the meta directly without reopening. For now we accept the
                 // over-default-fallback risk and recommend doctor's output.
                 const model = resolveModel(undefined);
-                const db = new EmbedDb({ file: embedFile, vaultRoot: vault.root, modelAlias: model.alias, dim: model.dim });
+                // v2.17.0 — pass through the quantization mode from CLI so the
+                // schema check matches what build-embeddings wrote. Default
+                // "f32" matches v2.16- behavior for users who don't set it.
+                const quantization = opts.quantizeEmbeddings ?? "f32";
+                const db = new EmbedDb({
+                    file: embedFile,
+                    vaultRoot: vault.root,
+                    modelAlias: model.alias,
+                    dim: model.dim,
+                    quantization
+                });
                 await db.open();
                 try {
                     const startMs = Date.now();
-                    const rows = db.getAllVectors();
-                    if (rows.length === 0) {
-                        process.stderr.write(`enquire: --use-hnsw passed but embed-db is empty; skipping HNSW build.\n`);
+                    // v2.16.0 — try to load from disk first if persistence is enabled.
+                    // Skip-rebuild path: ~50ms read vs ~25s build for 50K-chunk
+                    // vault when nothing changed since last serve. Staleness
+                    // detected via `EmbedDb.computeSignature()` mismatch.
+                    const persistFile = `${embedFile.replace(/\.embed\.db$/, "")}.hnsw`;
+                    const signature = db.computeSignature();
+                    const efOverride = opts.hnswEf ? parsePositiveInt(opts.hnswEf, "--hnsw-ef") : undefined;
+                    let loaded = null;
+                    if (opts.hnswPersist !== false) {
+                        const { loadHnswFromDisk } = await import("./hnsw.js");
+                        const loadResult = await loadHnswFromDisk(persistFile, signature);
+                        if (loadResult) {
+                            loaded = { index: loadResult.index, rowByLabel: loadResult.rowsByLabel };
+                            process.stderr.write(`enquire: HNSW index loaded from disk (${loadResult.index.size} vectors, dim=${loadResult.index.dim}, ${Date.now() - startMs}ms — signature matched)\n`);
+                        }
+                    }
+                    if (loaded) {
+                        hnswContext = {
+                            index: loaded.index,
+                            rowByLabel: loaded.rowByLabel,
+                            ...(efOverride !== undefined ? { ef: efOverride } : {})
+                        };
                     }
                     else {
-                        const { buildHnsw } = await import("./hnsw.js");
-                        const index = await buildHnsw(rows.map((r) => ({ label: r.label, vector: r.vector })), { dim: model.dim, maxElements: rows.length });
-                        const rowByLabel = new Map();
-                        for (const r of rows) {
-                            rowByLabel.set(r.label, {
-                                rel_path: r.rel_path,
-                                chunk_index: r.chunk_index,
-                                line_start: r.line_start,
-                                line_end: r.line_end,
-                                text_preview: r.text_preview,
-                                kind: r.kind
-                            });
+                        const rows = db.getAllVectors();
+                        if (rows.length === 0) {
+                            process.stderr.write(`enquire: --use-hnsw passed but embed-db is empty; skipping HNSW build.\n`);
+                        }
+                        else {
+                            const { buildHnsw } = await import("./hnsw.js");
+                            const index = await buildHnsw(rows.map((r) => ({ label: r.label, vector: r.vector })), { dim: model.dim, maxElements: rows.length });
+                            const rowByLabel = new Map();
+                            for (const r of rows) {
+                                rowByLabel.set(r.label, {
+                                    rel_path: r.rel_path,
+                                    chunk_index: r.chunk_index,
+                                    line_start: r.line_start,
+                                    line_end: r.line_end,
+                                    text_preview: r.text_preview,
+                                    kind: r.kind
+                                });
+                            }
+                            hnswContext = { index, rowByLabel, ...(efOverride !== undefined ? { ef: efOverride } : {}) };
+                            process.stderr.write(`enquire: HNSW index built (${rows.length} vectors, dim=${model.dim}, ${Date.now() - startMs}ms)\n`);
+                            // v2.16.0 — persist the freshly-built index for next serve start.
+                            if (opts.hnswPersist !== false) {
+                                try {
+                                    await index.saveTo(persistFile, rowByLabel, signature);
+                                    process.stderr.write(`enquire: HNSW index persisted to ${persistFile}.bin (+ .meta.json)\n`);
+                                }
+                                catch (err) {
+                                    // Non-fatal — persistence is an optimization. Log + continue.
+                                    process.stderr.write(`enquire: HNSW persist failed (continuing with in-memory index) — ${err instanceof Error ? err.message : String(err)}\n`);
+                                }
+                            }
                         }
-                        const efOverride = opts.hnswEf ? parsePositiveInt(opts.hnswEf, "--hnsw-ef") : undefined;
-                        hnswContext = { index, rowByLabel, ...(efOverride !== undefined ? { ef: efOverride } : {}) };
-                        process.stderr.write(`enquire: HNSW index built (${rows.length} vectors, dim=${model.dim}, ${Date.now() - startMs}ms)\n`);
                     }
                 }
                 finally {
@@ -2461,6 +2532,23 @@ function parsePositiveInt(raw, flag) {
     }
     return n;
 }
+/**
+ * v2.17.0 — validate a `--quantize-embeddings <mode>` value. Accepts the
+ * canonical `"f32"` / `"int8"` plus a few user-friendly aliases (`"none"`
+ * for f32; `"q8"` for int8). Anything else throws with the exact list of
+ * accepted values so the user can fix the typo immediately.
+ */
+function parseQuantizationMode(raw) {
+    if (raw === undefined)
+        return undefined;
+    const norm = raw.trim().toLowerCase();
+    if (norm === "" || norm === "f32" || norm === "float32" || norm === "none")
+        return "f32";
+    if (norm === "int8" || norm === "i8" || norm === "q8")
+        return "int8";
+    throw new Error(`--quantize-embeddings must be "f32" or "int8" (got "${raw}"). ` +
+        `Aliases: "none"/"float32" → f32, "q8"/"i8" → int8.`);
+}
 function encodeNotePath(relPath) {
     return relPath.split(path.sep).map(encodeURIComponent).join("/");
 }
@@ -2494,5 +2582,5 @@ if (isCliEntry) {
         process.exit(1);
     });
 }
-export { main, parsePositiveInt, startServer };
+export { main, parsePositiveInt, parseQuantizationMode, startServer };
 //# sourceMappingURL=index.js.map