better-web-search-mcp 0.3.0 → 0.4.1

package/README.md

@@ -72,15 +72,85 @@ Only escalates when needed. Results carry confidence scores so you know what you
 
 **Self-learning cache.** The first visit to a domain takes the full path. The second visit skips straight to what worked. Domain profiles and API patterns are remembered.
 
+## Token efficiency
+
+Doing the web work inside the server, rather than handing pages to the model,
+is the point. The repository ships the benchmark that checks it.
+
+Same questions, same pages, two workflows:
+
+- **baseline** — the agent drives the tools: one `web_search`, then
+  `web_extract` on each of the top 5 results. Every page lands in its context.
+- **research** — one `web_research` call, returning cited passages.
+
+12 questions across definition, technical, comparison, troubleshooting and
+research categories, run 2026-09-02 against the live web with the browser tier
+disabled:
+
+| | Baseline | `web_research` |
+|---|---|---|
+| Payload | 820,229 chars | **110,973 chars** |
+| Tokens (est. at 4 chars) | ~205,000 | **~28,000** |
+| Wall clock | 137.3s | **49.6s** |
+
+**86.5% less text overall, 83.8% median**, per-question range 56.8% to 93.4%.
+
+Reproduce it yourself:
+
+```bash
+npm run build && npm run bench
+```
+
+Honest caveats, because these are the ones that matter:
+
+- Payload is counted in **characters** — exact and tokenizer-independent. The
+  token column is an estimate at 4 chars/token for orientation; the *ratio* is
+  the real result and barely moves between tokenizers.
+- **This measures payload, not answer quality.** Deciding whether the retained
+  passages still support a correct answer needs a judge model, which this
+  harness does not have. A smaller payload is only a win if the right text
+  survived; `query_term_coverage` is reported per question as a weak proxy and
+  is deliberately not folded into the headline number.
+- Reduction tracks how much page text there was. The 56.8% case had a small
+  baseline (~20k chars), where the fixed cost of the response's source list
+  dominates.
+- Live-web runs are not bit-reproducible: pages change and DuckDuckGo
+  rate-limits scrapers.
+- No comparison against Exa, Tavily's answer endpoint, or hosted web search —
+  they cannot be run under identical conditions on the same pages.
+
+Method and raw per-question results: [`benchmarks/`](benchmarks/).
+
+## Security
+
+Page content is attacker-controlled and URLs arrive from the calling agent, so
+both are treated as untrusted.
+
+- **SSRF guard.** Non-HTTP schemes are refused; hostnames are resolved and
+  private, loopback, link-local, CGNAT, multicast and reserved addresses
+  rejected — including IPv4-mapped IPv6 forms like `::ffff:127.0.0.1`.
+  Redirects are followed manually so *every hop* is re-checked, capped at 5.
+- **Prompt-injection screening.** Every `web_extract` result carries a
+  `security` block marking content untrusted and flagging text that tries to
+  issue instructions to an agent, with the matched pattern and its offset.
+  Content is never rewritten, so extraction stays faithful to the source.
+- **Source independence.** Pages are clustered by content before citing, so a
+  wire story republished by five outlets counts as one account rather than five
+  corroborating sources.
+
+See [SECURITY.md](.github/SECURITY.md) for the full threat model.
+
 ## Project structure
 
 ```
 src/
   providers/      Search backends (DuckDuckGo, Brave, Tavily; SerpApi stubbed)
   extraction/     3-tier content pipeline (fetch, structured, browser)
-  ranking/        Deduplication, domain scoring, reranking
+  ranking/        Deduplication, domain scoring, reranking, passage selection,
+                  source-independence clustering
   tools/          MCP tool definitions (search, research, extract, find, news)
-  utils/          Config, caching, retry, query rewriting
+  utils/          Config, caching, retry, query rewriting, SSRF guard
+benchmarks/       Token-efficiency harness and results
 ```
 
 ## Configuration

package/dist/ranking/independence.d.ts

@@ -0,0 +1,70 @@
+/**
+ * Source independence via near-duplicate detection.
+ *
+ * URL deduplication only catches the same page twice. It does nothing about
+ * syndication: a wire story republished by five outlets, a vendor press
+ * release quoted verbatim, or an article and its own AMP variant all arrive as
+ * distinct URLs. Counting those as five sources is how a research tool reports
+ * strong agreement for what is really one claim from one author.
+ *
+ * Pages are compared by content shingles here, and near-duplicates are grouped
+ * into clusters with one primary per cluster. Callers can then say "three
+ * independent sources, two derivative" instead of "five sources", and citation
+ * selection can spread across clusters rather than across URLs.
+ *
+ * The comparison is Jaccard similarity over word shingles: deterministic,
+ * language-agnostic, and cheap at the page counts involved (research opens at
+ * most ten, so the quadratic comparison is a few dozen set intersections).
+ */
+/** Words per shingle. Long enough to be distinctive, short enough to survive edits. */
+export declare const SHINGLE_SIZE = 5;
+/** Jaccard similarity at or above which two documents are near-duplicates. */
+export declare const DUPLICATE_THRESHOLD = 0.4;
+/** A document participating in independence analysis. */
+export interface IndependenceInput {
+    /** The source URL, used to derive the registrable-ish host. */
+    url: string;
+    /** The extracted page content. */
+    content: string;
+}
+/** Where a document sits among its near-duplicates. */
+export interface IndependenceResult {
+    /** Index of the cluster this document belongs to. */
+    cluster: number;
+    /**
+     * Whether this is the representative of its cluster. Exactly one document
+     * per cluster is primary; the rest are derivative copies of the same text.
+     */
+    primary: boolean;
+    /** How many other documents share this cluster. */
+    duplicates: number;
+}
+/**
+ * Build the set of word shingles for a document.
+ *
+ * Exported for testing; callers normally use {@link analyzeIndependence}.
+ */
+export declare function shingles(text: string, size?: number): Set<string>;
+/** Jaccard similarity of two sets: |intersection| / |union|. */
+export declare function jaccard(a: ReadonlySet<string>, b: ReadonlySet<string>): number;
+/** The host of a URL, lowercased and stripped of a leading `www.`. */
+export declare function hostOf(url: string): string;
+/**
+ * Group documents into clusters of near-duplicates.
+ *
+ * Two documents join the same cluster when their shingle sets reach
+ * {@link DUPLICATE_THRESHOLD}, or when they share a host — the same site
+ * publishing two pages on a topic is not two independent confirmations.
+ * Clustering is transitive via union-find, so A~B and B~C puts all three
+ * together even when A and C alone fall below the threshold.
+ *
+ * The first document in each cluster is its primary, which keeps the incoming
+ * relevance order meaningful: callers rank before calling this.
+ *
+ * @param docs Documents in the caller's preferred (usually ranked) order.
+ * @returns One result per input document, in the same order.
+ */
+export declare function analyzeIndependence(docs: readonly IndependenceInput[]): IndependenceResult[];
+/** How many distinct clusters a set of independence results spans. */
+export declare function countIndependent(results: readonly IndependenceResult[]): number;
+//# sourceMappingURL=independence.d.ts.map

package/dist/ranking/independence.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"independence.d.ts","sourceRoot":"","sources":["../../src/ranking/independence.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;GAiBG;AAEH,uFAAuF;AACvF,eAAO,MAAM,YAAY,IAAI,CAAC;AAE9B,8EAA8E;AAC9E,eAAO,MAAM,mBAAmB,MAAM,CAAC;AAKvC,yDAAyD;AACzD,MAAM,WAAW,iBAAiB;IAChC,+DAA+D;IAC/D,GAAG,EAAE,MAAM,CAAC;IACZ,kCAAkC;IAClC,OAAO,EAAE,MAAM,CAAC;CACjB;AAED,uDAAuD;AACvD,MAAM,WAAW,kBAAkB;IACjC,qDAAqD;IACrD,OAAO,EAAE,MAAM,CAAC;IAChB;;;OAGG;IACH,OAAO,EAAE,OAAO,CAAC;IACjB,mDAAmD;IACnD,UAAU,EAAE,MAAM,CAAC;CACpB;AAOD;;;;GAIG;AACH,wBAAgB,QAAQ,CAAC,IAAI,EAAE,MAAM,EAAE,IAAI,SAAe,GAAG,GAAG,CAAC,MAAM,CAAC,CAevE;AAED,gEAAgE;AAChE,wBAAgB,OAAO,CAAC,CAAC,EAAE,WAAW,CAAC,MAAM,CAAC,EAAE,CAAC,EAAE,WAAW,CAAC,MAAM,CAAC,GAAG,MAAM,CAa9E;AAED,sEAAsE;AACtE,wBAAgB,MAAM,CAAC,GAAG,EAAE,MAAM,GAAG,MAAM,CAM1C;AAED;;;;;;;;;;;;;;GAcG;AACH,wBAAgB,mBAAmB,CACjC,IAAI,EAAE,SAAS,iBAAiB,EAAE,GACjC,kBAAkB,EAAE,CA0EtB;AAED,sEAAsE;AACtE,wBAAgB,gBAAgB,CAC9B,OAAO,EAAE,SAAS,kBAAkB,EAAE,GACrC,MAAM,CAER"}

package/dist/ranking/independence.js

@@ -0,0 +1,158 @@
+/**
+ * Source independence via near-duplicate detection.
+ *
+ * URL deduplication only catches the same page twice. It does nothing about
+ * syndication: a wire story republished by five outlets, a vendor press
+ * release quoted verbatim, or an article and its own AMP variant all arrive as
+ * distinct URLs. Counting those as five sources is how a research tool reports
+ * strong agreement for what is really one claim from one author.
+ *
+ * Pages are compared by content shingles here, and near-duplicates are grouped
+ * into clusters with one primary per cluster. Callers can then say "three
+ * independent sources, two derivative" instead of "five sources", and citation
+ * selection can spread across clusters rather than across URLs.
+ *
+ * The comparison is Jaccard similarity over word shingles: deterministic,
+ * language-agnostic, and cheap at the page counts involved (research opens at
+ * most ten, so the quadratic comparison is a few dozen set intersections).
+ */
+/** Words per shingle. Long enough to be distinctive, short enough to survive edits. */
+export const SHINGLE_SIZE = 5;
+/** Jaccard similarity at or above which two documents are near-duplicates. */
+export const DUPLICATE_THRESHOLD = 0.4;
+/** Minimum tokens a document needs before similarity means anything. */
+const MIN_TOKENS = SHINGLE_SIZE * 4;
+/** Lowercase word tokens used to build shingles. */
+function tokens(text) {
+    return text.toLowerCase().match(/[a-z0-9]+/g) ?? [];
+}
+/**
+ * Build the set of word shingles for a document.
+ *
+ * Exported for testing; callers normally use {@link analyzeIndependence}.
+ */
+export function shingles(text, size = SHINGLE_SIZE) {
+    const words = tokens(text);
+    const set = new Set();
+    if (words.length < size) {
+        // Too short to shingle: fall back to the whole token string so identical
+        // short documents still match each other.
+        if (words.length > 0) {
+            set.add(words.join(" "));
+        }
+        return set;
+    }
+    for (let i = 0; i + size <= words.length; i += 1) {
+        set.add(words.slice(i, i + size).join(" "));
+    }
+    return set;
+}
+/** Jaccard similarity of two sets: |intersection| / |union|. */
+export function jaccard(a, b) {
+    if (a.size === 0 || b.size === 0) {
+        return 0;
+    }
+    const [small, large] = a.size <= b.size ? [a, b] : [b, a];
+    let intersection = 0;
+    for (const value of small) {
+        if (large.has(value)) {
+            intersection += 1;
+        }
+    }
+    const union = a.size + b.size - intersection;
+    return union === 0 ? 0 : intersection / union;
+}
+/** The host of a URL, lowercased and stripped of a leading `www.`. */
+export function hostOf(url) {
+    try {
+        return new URL(url).hostname.toLowerCase().replace(/^www\./, "");
+    }
+    catch {
+        return "";
+    }
+}
+/**
+ * Group documents into clusters of near-duplicates.
+ *
+ * Two documents join the same cluster when their shingle sets reach
+ * {@link DUPLICATE_THRESHOLD}, or when they share a host — the same site
+ * publishing two pages on a topic is not two independent confirmations.
+ * Clustering is transitive via union-find, so A~B and B~C puts all three
+ * together even when A and C alone fall below the threshold.
+ *
+ * The first document in each cluster is its primary, which keeps the incoming
+ * relevance order meaningful: callers rank before calling this.
+ *
+ * @param docs Documents in the caller's preferred (usually ranked) order.
+ * @returns One result per input document, in the same order.
+ */
+export function analyzeIndependence(docs) {
+    const parent = docs.map((_, index) => index);
+    const find = (index) => {
+        let root = index;
+        while (parent[root] !== root) {
+            root = parent[root] ?? root;
+        }
+        // Path compression keeps repeated lookups cheap.
+        let cursor = index;
+        while (parent[cursor] !== root) {
+            const next = parent[cursor] ?? root;
+            parent[cursor] = root;
+            cursor = next;
+        }
+        return root;
+    };
+    const union = (a, b) => {
+        const rootA = find(a);
+        const rootB = find(b);
+        if (rootA !== rootB) {
+            // Always attach to the lower index so the earliest document wins.
+            const [keep, drop] = rootA < rootB ? [rootA, rootB] : [rootB, rootA];
+            parent[drop] = keep;
+        }
+    };
+    const fingerprints = docs.map((doc) => shingles(doc.content));
+    const hosts = docs.map((doc) => hostOf(doc.url));
+    const lengths = docs.map((doc) => tokens(doc.content).length);
+    for (let i = 0; i < docs.length; i += 1) {
+        for (let j = i + 1; j < docs.length; j += 1) {
+            const sameHost = hosts[i] !== "" && hosts[i] === hosts[j];
+            if (sameHost) {
+                union(i, j);
+                continue;
+            }
+            // Documents too short to fingerprint are left independent rather than
+            // collapsed on a coincidental match.
+            if ((lengths[i] ?? 0) < MIN_TOKENS || (lengths[j] ?? 0) < MIN_TOKENS) {
+                continue;
+            }
+            const similarity = jaccard(fingerprints[i] ?? new Set(), fingerprints[j] ?? new Set());
+            if (similarity >= DUPLICATE_THRESHOLD) {
+                union(i, j);
+            }
+        }
+    }
+    // Number clusters by first appearance so output is stable and readable.
+    const clusterIds = new Map();
+    const members = new Map();
+    const roots = docs.map((_, index) => find(index));
+    for (const root of roots) {
+        if (!clusterIds.has(root)) {
+            clusterIds.set(root, clusterIds.size);
+        }
+        members.set(root, (members.get(root) ?? 0) + 1);
+    }
+    return docs.map((_, index) => {
+        const root = roots[index] ?? index;
+        return {
+            cluster: clusterIds.get(root) ?? 0,
+            primary: root === index,
+            duplicates: (members.get(root) ?? 1) - 1,
+        };
+    });
+}
+/** How many distinct clusters a set of independence results spans. */
+export function countIndependent(results) {
+    return new Set(results.map((result) => result.cluster)).size;
+}
+//# sourceMappingURL=independence.js.map

package/dist/ranking/independence.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"independence.js","sourceRoot":"","sources":["../../src/ranking/independence.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;GAiBG;AAEH,uFAAuF;AACvF,MAAM,CAAC,MAAM,YAAY,GAAG,CAAC,CAAC;AAE9B,8EAA8E;AAC9E,MAAM,CAAC,MAAM,mBAAmB,GAAG,GAAG,CAAC;AAEvC,wEAAwE;AACxE,MAAM,UAAU,GAAG,YAAY,GAAG,CAAC,CAAC;AAuBpC,oDAAoD;AACpD,SAAS,MAAM,CAAC,IAAY;IAC1B,OAAO,IAAI,CAAC,WAAW,EAAE,CAAC,KAAK,CAAC,YAAY,CAAC,IAAI,EAAE,CAAC;AACtD,CAAC;AAED;;;;GAIG;AACH,MAAM,UAAU,QAAQ,CAAC,IAAY,EAAE,IAAI,GAAG,YAAY;IACxD,MAAM,KAAK,GAAG,MAAM,CAAC,IAAI,CAAC,CAAC;IAC3B,MAAM,GAAG,GAAG,IAAI,GAAG,EAAU,CAAC;IAC9B,IAAI,KAAK,CAAC,MAAM,GAAG,IAAI,EAAE,CAAC;QACxB,yEAAyE;QACzE,0CAA0C;QAC1C,IAAI,KAAK,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;YACrB,GAAG,CAAC,GAAG,CAAC,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC;QAC3B,CAAC;QACD,OAAO,GAAG,CAAC;IACb,CAAC;IACD,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,IAAI,IAAI,KAAK,CAAC,MAAM,EAAE,CAAC,IAAI,CAAC,EAAE,CAAC;QACjD,GAAG,CAAC,GAAG,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,GAAG,IAAI,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC;IAC9C,CAAC;IACD,OAAO,GAAG,CAAC;AACb,CAAC;AAED,gEAAgE;AAChE,MAAM,UAAU,OAAO,CAAC,CAAsB,EAAE,CAAsB;IACpE,IAAI,CAAC,CAAC,IAAI,KAAK,CAAC,IAAI,CAAC,CAAC,IAAI,KAAK,CAAC,EAAE,CAAC;QACjC,OAAO,CAAC,CAAC;IACX,CAAC;IACD,MAAM,CAAC,KAAK,EAAE,KAAK,CAAC,GAAG,CAAC,CAAC,IAAI,IAAI,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC;IAC1D,IAAI,YAAY,GAAG,CAAC,CAAC;IACrB,KAAK,MAAM,KAAK,IAAI,KAAK,EAAE,CAAC;QAC1B,IAAI,KAAK,CAAC,GAAG,CAAC,KAAK,CAAC,EAAE,CAAC;YACrB,YAAY,IAAI,CAAC,CAAC;QACpB,CAAC;IACH,CAAC;IACD,MAAM,KAAK,GAAG,CAAC,CAAC,IAAI,GAAG,CAAC,CAAC,IAAI,GAAG,YAAY,CAAC;IAC7C,OAAO,KAAK,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,YAAY,GAAG,KAAK,CAAC;AAChD,CAAC;AAED,sEAAsE;AACtE,MAAM,UAAU,MAAM,CAAC,GAAW;IAChC,IAAI,CAAC;QACH,OAAO,IAAI,GAAG,CAAC,GAAG,CAAC,CAAC,QAAQ,CAAC,WAAW,EAAE,CAAC,OAAO,CAAC,QAAQ,EAAE,EAAE,CAAC,CAAC;IACnE,CAAC;IAAC,MAAM,CAAC;QACP,OAAO,EAAE,CAAC;IACZ,CAAC;AACH,CAAC;AAED;;;;;;;;;;;;;;GAcG;AACH,MAAM,UAAU,mBAAmB,CACjC,IAAkC;IAElC,MAAM,MAAM,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,CAAC;IAE7C,MAAM,IAAI,GAAG,CAAC,KAAa,EAAU,EAAE;QACrC,IAAI,IAAI,GAAG,KAAK,CAAC;QACjB,OAAO,MAAM,CAAC,IAAI,CAAC,KAAK,IAAI,EAAE,CAAC;YAC7B,IAAI,GAAG,MAAM,CAAC,IAAI,CAAC,IAAI,IAAI,CAAC;QAC9B,CAAC;QACD,iDAAiD;QACjD,IAAI,MAAM,GAAG,KAAK,CAAC;QACnB,OAAO,MAAM,CAAC,MAAM,CAAC,KAAK,IAAI,EAAE,CAAC;YAC/B,MAAM,IAAI,GAAG,MAAM,CAAC,MAAM,CAAC,IAAI,IAAI,CAAC;YACpC,MAAM,CAAC,MAAM,CAAC,GAAG,IAAI,CAAC;YACtB,MAAM,GAAG,IAAI,CAAC;QAChB,CAAC;QACD,OAAO,IAAI,CAAC;IACd,CAAC,CAAC;IAEF,MAAM,KAAK,GAAG,CAAC,CAAS,EAAE,CAAS,EAAQ,EAAE;QAC3C,MAAM,KAAK,GAAG,IAAI,CAAC,CAAC,CAAC,CAAC;QACtB,MAAM,KAAK,GAAG,IAAI,CAAC,CAAC,CAAC,CAAC;QACtB,IAAI,KAAK,KAAK,KAAK,EAAE,CAAC;YACpB,kEAAkE;YAClE,MAAM,CAAC,IAAI,EAAE,IAAI,CAAC,GAAG,KAAK,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC,KAAK,EAAE,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,KAAK,EAAE,KAAK,CAAC,CAAC;YACrE,MAAM,CAAC,IAAI,CAAC,GAAG,IAAI,CAAC;QACtB,CAAC;IACH,CAAC,CAAC;IAEF,MAAM,YAAY,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,QAAQ,CAAC,GAAG,CAAC,OAAO,CAAC,CAAC,CAAC;IAC9D,MAAM,KAAK,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,MAAM,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC,CAAC;IACjD,MAAM,OAAO,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,MAAM,CAAC,GAAG,CAAC,OAAO,CAAC,CAAC,MAAM,CAAC,CAAC;IAE9D,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,IAAI,CAAC,MAAM,EAAE,CAAC,IAAI,CAAC,EAAE,CAAC;QACxC,KAAK,IAAI,CAAC,GAAG,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,IAAI,CAAC,MAAM,EAAE,CAAC,IAAI,CAAC,EAAE,CAAC;YAC5C,MAAM,QAAQ,GACZ,KAAK,CAAC,CAAC,CAAC,KAAK,EAAE,IAAI,KAAK,CAAC,CAAC,CAAC,KAAK,KAAK,CAAC,CAAC,CAAC,CAAC;YAC3C,IAAI,QAAQ,EAAE,CAAC;gBACb,KAAK,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC;gBACZ,SAAS;YACX,CAAC;YACD,sEAAsE;YACtE,qCAAqC;YACrC,IAAI,CAAC,OAAO,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,GAAG,UAAU,IAAI,CAAC,OAAO,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,GAAG,UAAU,EAAE,CAAC;gBACrE,SAAS;YACX,CAAC;YACD,MAAM,UAAU,GAAG,OAAO,CACxB,YAAY,CAAC,CAAC,CAAC,IAAI,IAAI,GAAG,EAAE,EAC5B,YAAY,CAAC,CAAC,CAAC,IAAI,IAAI,GAAG,EAAE,CAC7B,CAAC;YACF,IAAI,UAAU,IAAI,mBAAmB,EAAE,CAAC;gBACtC,KAAK,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC;YACd,CAAC;QACH,CAAC;IACH,CAAC;IAED,wEAAwE;IACxE,MAAM,UAAU,GAAG,IAAI,GAAG,EAAkB,CAAC;IAC7C,MAAM,OAAO,GAAG,IAAI,GAAG,EAAkB,CAAC;IAC1C,MAAM,KAAK,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,KAAK,EAAE,EAAE,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC;IAClD,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;QACzB,IAAI,CAAC,UAAU,CAAC,GAAG,CAAC,IAAI,CAAC,EAAE,CAAC;YAC1B,UAAU,CAAC,GAAG,CAAC,IAAI,EAAE,UAAU,CAAC,IAAI,CAAC,CAAC;QACxC,CAAC;QACD,OAAO,CAAC,GAAG,CAAC,IAAI,EAAE,CAAC,OAAO,CAAC,GAAG,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC;IAClD,CAAC;IAED,OAAO,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,KAAK,EAAE,EAAE;QAC3B,MAAM,IAAI,GAAG,KAAK,CAAC,KAAK,CAAC,IAAI,KAAK,CAAC;QACnC,OAAO;YACL,OAAO,EAAE,UAAU,CAAC,GAAG,CAAC,IAAI,CAAC,IAAI,CAAC;YAClC,OAAO,EAAE,IAAI,KAAK,KAAK;YACvB,UAAU,EAAE,CAAC,OAAO,CAAC,GAAG,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,GAAG,CAAC;SACzC,CAAC;IACJ,CAAC,CAAC,CAAC;AACL,CAAC;AAED,sEAAsE;AACtE,MAAM,UAAU,gBAAgB,CAC9B,OAAsC;IAEtC,OAAO,IAAI,GAAG,CAAC,OAAO,CAAC,GAAG,CAAC,CAAC,MAAM,EAAE,EAAE,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC,CAAC,IAAI,CAAC;AAC/D,CAAC"}

package/dist/ranking/passages.d.ts

@@ -28,6 +28,16 @@ export interface Passage {
 }
 /** Target passage size in characters; paragraphs are packed up to this. */
 export declare const TARGET_PASSAGE_CHARS = 500;
+/**
+ * Hard ceiling on a single passage, in characters.
+ *
+ * Paragraph splitting alone is not enough: plenty of extracted pages contain
+ * no blank lines at all, which made the whole page one passage. A cited
+ * "passage" was then the entire document, and `web_research` returned more
+ * text than reading the pages directly would have. Anything longer than this
+ * is windowed on sentence boundaries.
+ */
+export declare const MAX_PASSAGE_CHARS = 1200;
 /** Lowercase alphanumeric tokens, stop words removed. */
 export declare function tokenize(text: string): string[];
 /**

package/dist/ranking/passages.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"passages.d.ts","sourceRoot":"","sources":["../../src/ranking/passages.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;GAgBG;AAEH,+CAA+C;AAC/C,MAAM,WAAW,OAAO;IACtB,iCAAiC;IACjC,IAAI,EAAE,MAAM,CAAC;IACb,uEAAuE;IACvE,KAAK,EAAE,MAAM,CAAC;IACd,qEAAqE;IACrE,GAAG,EAAE,MAAM,CAAC;IACZ,6DAA6D;IAC7D,KAAK,EAAE,MAAM,CAAC;CACf;AAMD,2EAA2E;AAC3E,eAAO,MAAM,oBAAoB,MAAM,CAAC;AAUxC,yDAAyD;AACzD,wBAAgB,QAAQ,CAAC,IAAI,EAAE,MAAM,GAAG,MAAM,EAAE,CAG/C;AAED;;;;;;;GAOG;AACH,wBAAgB,aAAa,CAAC,OAAO,EAAE,MAAM,GAAG,OAAO,EAAE,CAuCxD;AAED;;;;;;GAMG;AACH,wBAAgB,YAAY,CAC1B,QAAQ,EAAE,SAAS,OAAO,EAAE,EAC5B,KAAK,EAAE,MAAM,GACZ,OAAO,EAAE,CAqDX;AAED;;;;;;GAMG;AACH,wBAAgB,cAAc,CAC5B,OAAO,EAAE,MAAM,EACf,KAAK,EAAE,MAAM,EACb,KAAK,SAAI,GACR,OAAO,EAAE,CAEX"}
+{"version":3,"file":"passages.d.ts","sourceRoot":"","sources":["../../src/ranking/passages.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;GAgBG;AAEH,+CAA+C;AAC/C,MAAM,WAAW,OAAO;IACtB,iCAAiC;IACjC,IAAI,EAAE,MAAM,CAAC;IACb,uEAAuE;IACvE,KAAK,EAAE,MAAM,CAAC;IACd,qEAAqE;IACrE,GAAG,EAAE,MAAM,CAAC;IACZ,6DAA6D;IAC7D,KAAK,EAAE,MAAM,CAAC;CACf;AAMD,2EAA2E;AAC3E,eAAO,MAAM,oBAAoB,MAAM,CAAC;AAExC;;;;;;;;GAQG;AACH,eAAO,MAAM,iBAAiB,OAAQ,CAAC;AAUvC,yDAAyD;AACzD,wBAAgB,QAAQ,CAAC,IAAI,EAAE,MAAM,GAAG,MAAM,EAAE,CAG/C;AAED;;;;;;;GAOG;AACH,wBAAgB,aAAa,CAAC,OAAO,EAAE,MAAM,GAAG,OAAO,EAAE,CAyExD;AAED;;;;;;GAMG;AACH,wBAAgB,YAAY,CAC1B,QAAQ,EAAE,SAAS,OAAO,EAAE,EAC5B,KAAK,EAAE,MAAM,GACZ,OAAO,EAAE,CAqDX;AAED;;;;;;GAMG;AACH,wBAAgB,cAAc,CAC5B,OAAO,EAAE,MAAM,EACf,KAAK,EAAE,MAAM,EACb,KAAK,SAAI,GACR,OAAO,EAAE,CAEX"}

package/dist/ranking/passages.js

@@ -20,6 +20,16 @@ const K1 = 1.5;
 const B = 0.75;
 /** Target passage size in characters; paragraphs are packed up to this. */
 export const TARGET_PASSAGE_CHARS = 500;
+/**
+ * Hard ceiling on a single passage, in characters.
+ *
+ * Paragraph splitting alone is not enough: plenty of extracted pages contain
+ * no blank lines at all, which made the whole page one passage. A cited
+ * "passage" was then the entire document, and `web_research` returned more
+ * text than reading the pages directly would have. Anything longer than this
+ * is windowed on sentence boundaries.
+ */
+export const MAX_PASSAGE_CHARS = 1_200;
 /** Words carrying no retrieval signal, skipped when scoring. */
 const STOP_WORDS = new Set([
     "a", "an", "and", "are", "as", "at", "be", "but", "by", "for", "from",
@@ -46,11 +56,44 @@ export function splitPassages(content) {
     let bufferStart = -1;
     let bufferEnd = -1;
     let buffer = "";
-    const flush = () => {
-        const text = buffer.trim();
-        if (text !== "") {
-            passages.push({ text, start: bufferStart, end: bufferEnd, score: 0 });
+    const emit = (text, start, end) => {
+        if (text.trim() === "") {
+            return;
+        }
+        if (text.length <= MAX_PASSAGE_CHARS) {
+            passages.push({ text: text.trim(), start, end, score: 0 });
+            return;
+        }
+        // Window an over-long block on sentence boundaries, keeping offsets
+        // accurate by tracking position in the original string.
+        let windowStart = 0;
+        while (windowStart < text.length) {
+            let windowEnd = Math.min(windowStart + MAX_PASSAGE_CHARS, text.length);
+            if (windowEnd < text.length) {
+                // Prefer the last sentence end inside the window; fall back to a
+                // space so words are never cut in half.
+                const slice = text.slice(windowStart, windowEnd);
+                const sentence = slice.lastIndexOf(". ");
+                const space = slice.lastIndexOf(" ");
+                const cut = sentence > MAX_PASSAGE_CHARS / 2 ? sentence + 1 : space;
+                if (cut > 0) {
+                    windowEnd = windowStart + cut;
+                }
+            }
+            const chunk = text.slice(windowStart, windowEnd);
+            if (chunk.trim() !== "") {
+                passages.push({
+                    text: chunk.trim(),
+                    start: start + windowStart,
+                    end: start + windowEnd,
+                    score: 0,
+                });
+            }
+            windowStart = windowEnd;
         }
+    };
+    const flush = () => {
+        emit(buffer, bufferStart, bufferEnd);
         buffer = "";
         bufferStart = -1;
         bufferEnd = -1;

package/dist/ranking/passages.js.map

@@ -1 +1 @@
-{"version":3,"file":"passages.js","sourceRoot":"","sources":["../../src/ranking/passages.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;GAgBG;AAcH,oDAAoD;AACpD,MAAM,EAAE,GAAG,GAAG,CAAC;AACf,MAAM,CAAC,GAAG,IAAI,CAAC;AAEf,2EAA2E;AAC3E,MAAM,CAAC,MAAM,oBAAoB,GAAG,GAAG,CAAC;AAExC,gEAAgE;AAChE,MAAM,UAAU,GAAG,IAAI,GAAG,CAAC;IACzB,GAAG,EAAE,IAAI,EAAE,KAAK,EAAE,KAAK,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,KAAK,EAAE,IAAI,EAAE,KAAK,EAAE,MAAM;IACrE,KAAK,EAAE,MAAM,EAAE,KAAK,EAAE,GAAG,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,KAAK,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,GAAG;IACzE,MAAM,EAAE,KAAK,EAAE,MAAM,EAAE,IAAI,EAAE,KAAK,EAAE,MAAM,EAAE,MAAM,EAAE,MAAM,EAAE,OAAO,EAAE,OAAO;IAC5E,KAAK,EAAE,KAAK,EAAE,MAAM,EAAE,MAAM,EAAE,KAAK,EAAE,MAAM;CAC5C,CAAC,CAAC;AAEH,yDAAyD;AACzD,MAAM,UAAU,QAAQ,CAAC,IAAY;IACnC,MAAM,MAAM,GAAG,IAAI,CAAC,WAAW,EAAE,CAAC,KAAK,CAAC,aAAa,CAAC,IAAI,EAAE,CAAC;IAC7D,OAAO,MAAM,CAAC,MAAM,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,MAAM,GAAG,CAAC,IAAI,CAAC,UAAU,CAAC,GAAG,CAAC,KAAK,CAAC,CAAC,CAAC;AAC9E,CAAC;AAED;;;;;;;GAOG;AACH,MAAM,UAAU,aAAa,CAAC,OAAe;IAC3C,MAAM,QAAQ,GAAc,EAAE,CAAC;IAC/B,MAAM,WAAW,GAAG,+BAA+B,CAAC;IAEpD,IAAI,WAAW,GAAG,CAAC,CAAC,CAAC;IACrB,IAAI,SAAS,GAAG,CAAC,CAAC,CAAC;IACnB,IAAI,MAAM,GAAG,EAAE,CAAC;IAEhB,MAAM,KAAK,GAAG,GAAS,EAAE;QACvB,MAAM,IAAI,GAAG,MAAM,CAAC,IAAI,EAAE,CAAC;QAC3B,IAAI,IAAI,KAAK,EAAE,EAAE,CAAC;YAChB,QAAQ,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,KAAK,EAAE,WAAW,EAAE,GAAG,EAAE,SAAS,EAAE,KAAK,EAAE,CAAC,EAAE,CAAC,CAAC;QACxE,CAAC;QACD,MAAM,GAAG,EAAE,CAAC;QACZ,WAAW,GAAG,CAAC,CAAC,CAAC;QACjB,SAAS,GAAG,CAAC,CAAC,CAAC;IACjB,CAAC,CAAC;IAEF,KAAK,MAAM,KAAK,IAAI,OAAO,CAAC,QAAQ,CAAC,WAAW,CAAC,EAAE,CAAC;QAClD,MAAM,SAAS,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC;QAC3B,MAAM,KAAK,GAAG,KAAK,CAAC,KAAK,IAAI,CAAC,CAAC;QAC/B,IAAI,SAAS,CAAC,IAAI,EAAE,KAAK,EAAE,EAAE,CAAC;YAC5B,SAAS;QACX,CAAC;QACD,IAAI,MAAM,KAAK,EAAE,EAAE,CAAC;YAClB,WAAW,GAAG,KAAK,CAAC;QACtB,CAAC;aAAM,IAAI,MAAM,CAAC,MAAM,GAAG,SAAS,CAAC,MAAM,GAAG,oBAAoB,EAAE,CAAC;YACnE,KAAK,EAAE,CAAC;YACR,WAAW,GAAG,KAAK,CAAC;QACtB,CAAC;QACD,MAAM,GAAG,MAAM,KAAK,EAAE,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,GAAG,MAAM,OAAO,SAAS,EAAE,CAAC;QACjE,SAAS,GAAG,KAAK,GAAG,SAAS,CAAC,MAAM,CAAC;QACrC,IAAI,MAAM,CAAC,MAAM,IAAI,oBAAoB,EAAE,CAAC;YAC1C,KAAK,EAAE,CAAC;QACV,CAAC;IACH,CAAC;IACD,KAAK,EAAE,CAAC;IAER,OAAO,QAAQ,CAAC;AAClB,CAAC;AAED;;;;;;GAMG;AACH,MAAM,UAAU,YAAY,CAC1B,QAA4B,EAC5B,KAAa;IAEb,MAAM,UAAU,GAAG,QAAQ,CAAC,KAAK,CAAC,CAAC;IACnC,IAAI,UAAU,CAAC,MAAM,KAAK,CAAC,IAAI,QAAQ,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;QACrD,OAAO,EAAE,CAAC;IACZ,CAAC;IAED,MAAM,IAAI,GAAG,QAAQ,CAAC,GAAG,CAAC,CAAC,OAAO,EAAE,EAAE,CAAC,QAAQ,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC;IAC/D,MAAM,SAAS,GACb,IAAI,CAAC,MAAM,CAAC,CAAC,GAAG,EAAE,GAAG,EAAE,EAAE,CAAC,GAAG,GAAG,GAAG,CAAC,MAAM,EAAE,CAAC,CAAC,GAAG,CAAC,IAAI,CAAC,MAAM,IAAI,CAAC,CAAC,CAAC;IAEtE,mEAAmE;IACnE,MAAM,OAAO,GAAG,IAAI,GAAG,EAAkB,CAAC;IAC1C,KAAK,MAAM,IAAI,IAAI,IAAI,GAAG,CAAC,UAAU,CAAC,EAAE,CAAC;QACvC,IAAI,KAAK,GAAG,CAAC,CAAC;QACd,KAAK,MAAM,GAAG,IAAI,IAAI,EAAE,CAAC;YACvB,IAAI,GAAG,CAAC,QAAQ,CAAC,IAAI,CAAC,EAAE,CAAC;gBACvB,KAAK,IAAI,CAAC,CAAC;YACb,CAAC;QACH,CAAC;QACD,OAAO,CAAC,GAAG,CAAC,IAAI,EAAE,KAAK,CAAC,CAAC;IAC3B,CAAC;IAED,MAAM,MAAM,GAAc,EAAE,CAAC;IAC7B,QAAQ,CAAC,OAAO,CAAC,CAAC,OAAO,EAAE,KAAK,EAAE,EAAE;QAClC,MAAM,GAAG,GAAG,IAAI,CAAC,KAAK,CAAC,IAAI,EAAE,CAAC;QAC9B,IAAI,GAAG,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;YACrB,OAAO;QACT,CAAC;QACD,MAAM,MAAM,GAAG,IAAI,GAAG,EAAkB,CAAC;QACzC,KAAK,MAAM,KAAK,IAAI,GAAG,EAAE,CAAC;YACxB,MAAM,CAAC,GAAG,CAAC,KAAK,EAAE,CAAC,MAAM,CAAC,GAAG,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC;QAClD,CAAC;QAED,IAAI,KAAK,GAAG,CAAC,CAAC;QACd,KAAK,MAAM,IAAI,IAAI,IAAI,GAAG,CAAC,UAAU,CAAC,EAAE,CAAC;YACvC,MAAM,IAAI,GAAG,MAAM,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC;YAC9B,IAAI,IAAI,KAAK,SAAS,EAAE,CAAC;gBACvB,SAAS;YACX,CAAC;YACD,MAAM,CAAC,GAAG,OAAO,CAAC,GAAG,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;YACjC,mEAAmE;YACnE,MAAM,GAAG,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,GAAG,CAAC,IAAI,CAAC,MAAM,GAAG,CAAC,GAAG,GAAG,CAAC,GAAG,CAAC,CAAC,GAAG,GAAG,CAAC,CAAC,CAAC;YAC9D,MAAM,IAAI,GAAG,IAAI,GAAG,CAAC,EAAE,GAAG,CAAC,CAAC,CAAC;YAC7B,MAAM,KAAK,GAAG,IAAI,GAAG,EAAE,GAAG,CAAC,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC,GAAG,GAAG,CAAC,MAAM,CAAC,GAAG,CAAC,SAAS,IAAI,CAAC,CAAC,CAAC,CAAC;YACxE,KAAK,IAAI,GAAG,GAAG,CAAC,IAAI,GAAG,KAAK,CAAC,CAAC;QAChC,CAAC;QAED,IAAI,KAAK,GAAG,CAAC,EAAE,CAAC;YACd,MAAM,CAAC,IAAI,CAAC,EAAE,GAAG,OAAO,EAAE,KAAK,EAAE,CAAC,CAAC;QACrC,CAAC;IACH,CAAC,CAAC,CAAC;IAEH,OAAO,MAAM,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,KAAK,GAAG,CAAC,CAAC,KAAK,CAAC,CAAC;AAClD,CAAC;AAED;;;;;;GAMG;AACH,MAAM,UAAU,cAAc,CAC5B,OAAe,EACf,KAAa,EACb,KAAK,GAAG,CAAC;IAET,OAAO,YAAY,CAAC,aAAa,CAAC,OAAO,CAAC,EAAE,KAAK,CAAC,CAAC,KAAK,CAAC,CAAC,EAAE,KAAK,CAAC,CAAC;AACrE,CAAC"}
+{"version":3,"file":"passages.js","sourceRoot":"","sources":["../../src/ranking/passages.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;GAgBG;AAcH,oDAAoD;AACpD,MAAM,EAAE,GAAG,GAAG,CAAC;AACf,MAAM,CAAC,GAAG,IAAI,CAAC;AAEf,2EAA2E;AAC3E,MAAM,CAAC,MAAM,oBAAoB,GAAG,GAAG,CAAC;AAExC;;;;;;;;GAQG;AACH,MAAM,CAAC,MAAM,iBAAiB,GAAG,KAAK,CAAC;AAEvC,gEAAgE;AAChE,MAAM,UAAU,GAAG,IAAI,GAAG,CAAC;IACzB,GAAG,EAAE,IAAI,EAAE,KAAK,EAAE,KAAK,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,KAAK,EAAE,IAAI,EAAE,KAAK,EAAE,MAAM;IACrE,KAAK,EAAE,MAAM,EAAE,KAAK,EAAE,GAAG,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,KAAK,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,GAAG;IACzE,MAAM,EAAE,KAAK,EAAE,MAAM,EAAE,IAAI,EAAE,KAAK,EAAE,MAAM,EAAE,MAAM,EAAE,MAAM,EAAE,OAAO,EAAE,OAAO;IAC5E,KAAK,EAAE,KAAK,EAAE,MAAM,EAAE,MAAM,EAAE,KAAK,EAAE,MAAM;CAC5C,CAAC,CAAC;AAEH,yDAAyD;AACzD,MAAM,UAAU,QAAQ,CAAC,IAAY;IACnC,MAAM,MAAM,GAAG,IAAI,CAAC,WAAW,EAAE,CAAC,KAAK,CAAC,aAAa,CAAC,IAAI,EAAE,CAAC;IAC7D,OAAO,MAAM,CAAC,MAAM,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,MAAM,GAAG,CAAC,IAAI,CAAC,UAAU,CAAC,GAAG,CAAC,KAAK,CAAC,CAAC,CAAC;AAC9E,CAAC;AAED;;;;;;;GAOG;AACH,MAAM,UAAU,aAAa,CAAC,OAAe;IAC3C,MAAM,QAAQ,GAAc,EAAE,CAAC;IAC/B,MAAM,WAAW,GAAG,+BAA+B,CAAC;IAEpD,IAAI,WAAW,GAAG,CAAC,CAAC,CAAC;IACrB,IAAI,SAAS,GAAG,CAAC,CAAC,CAAC;IACnB,IAAI,MAAM,GAAG,EAAE,CAAC;IAEhB,MAAM,IAAI,GAAG,CAAC,IAAY,EAAE,KAAa,EAAE,GAAW,EAAQ,EAAE;QAC9D,IAAI,IAAI,CAAC,IAAI,EAAE,KAAK,EAAE,EAAE,CAAC;YACvB,OAAO;QACT,CAAC;QACD,IAAI,IAAI,CAAC,MAAM,IAAI,iBAAiB,EAAE,CAAC;YACrC,QAAQ,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,IAAI,CAAC,IAAI,EAAE,EAAE,KAAK,EAAE,GAAG,EAAE,KAAK,EAAE,CAAC,EAAE,CAAC,CAAC;YAC3D,OAAO;QACT,CAAC;QACD,oEAAoE;QACpE,wDAAwD;QACxD,IAAI,WAAW,GAAG,CAAC,CAAC;QACpB,OAAO,WAAW,GAAG,IAAI,CAAC,MAAM,EAAE,CAAC;YACjC,IAAI,SAAS,GAAG,IAAI,CAAC,GAAG,CAAC,WAAW,GAAG,iBAAiB,EAAE,IAAI,CAAC,MAAM,CAAC,CAAC;YACvE,IAAI,SAAS,GAAG,IAAI,CAAC,MAAM,EAAE,CAAC;gBAC5B,iEAAiE;gBACjE,wCAAwC;gBACxC,MAAM,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,WAAW,EAAE,SAAS,CAAC,CAAC;gBACjD,MAAM,QAAQ,GAAG,KAAK,CAAC,WAAW,CAAC,IAAI,CAAC,CAAC;gBACzC,MAAM,KAAK,GAAG,KAAK,CAAC,WAAW,CAAC,GAAG,CAAC,CAAC;gBACrC,MAAM,GAAG,GAAG,QAAQ,GAAG,iBAAiB,GAAG,CAAC,CAAC,CAAC,CAAC,QAAQ,GAAG,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC;gBACpE,IAAI,GAAG,GAAG,CAAC,EAAE,CAAC;oBACZ,SAAS,GAAG,WAAW,GAAG,GAAG,CAAC;gBAChC,CAAC;YACH,CAAC;YACD,MAAM,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,WAAW,EAAE,SAAS,CAAC,CAAC;YACjD,IAAI,KAAK,CAAC,IAAI,EAAE,KAAK,EAAE,EAAE,CAAC;gBACxB,QAAQ,CAAC,IAAI,CAAC;oBACZ,IAAI,EAAE,KAAK,CAAC,IAAI,EAAE;oBAClB,KAAK,EAAE,KAAK,GAAG,WAAW;oBAC1B,GAAG,EAAE,KAAK,GAAG,SAAS;oBACtB,KAAK,EAAE,CAAC;iBACT,CAAC,CAAC;YACL,CAAC;YACD,WAAW,GAAG,SAAS,CAAC;QAC1B,CAAC;IACH,CAAC,CAAC;IAEF,MAAM,KAAK,GAAG,GAAS,EAAE;QACvB,IAAI,CAAC,MAAM,EAAE,WAAW,EAAE,SAAS,CAAC,CAAC;QACrC,MAAM,GAAG,EAAE,CAAC;QACZ,WAAW,GAAG,CAAC,CAAC,CAAC;QACjB,SAAS,GAAG,CAAC,CAAC,CAAC;IACjB,CAAC,CAAC;IAEF,KAAK,MAAM,KAAK,IAAI,OAAO,CAAC,QAAQ,CAAC,WAAW,CAAC,EAAE,CAAC;QAClD,MAAM,SAAS,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC;QAC3B,MAAM,KAAK,GAAG,KAAK,CAAC,KAAK,IAAI,CAAC,CAAC;QAC/B,IAAI,SAAS,CAAC,IAAI,EAAE,KAAK,EAAE,EAAE,CAAC;YAC5B,SAAS;QACX,CAAC;QACD,IAAI,MAAM,KAAK,EAAE,EAAE,CAAC;YAClB,WAAW,GAAG,KAAK,CAAC;QACtB,CAAC;aAAM,IAAI,MAAM,CAAC,MAAM,GAAG,SAAS,CAAC,MAAM,GAAG,oBAAoB,EAAE,CAAC;YACnE,KAAK,EAAE,CAAC;YACR,WAAW,GAAG,KAAK,CAAC;QACtB,CAAC;QACD,MAAM,GAAG,MAAM,KAAK,EAAE,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,GAAG,MAAM,OAAO,SAAS,EAAE,CAAC;QACjE,SAAS,GAAG,KAAK,GAAG,SAAS,CAAC,MAAM,CAAC;QACrC,IAAI,MAAM,CAAC,MAAM,IAAI,oBAAoB,EAAE,CAAC;YAC1C,KAAK,EAAE,CAAC;QACV,CAAC;IACH,CAAC;IACD,KAAK,EAAE,CAAC;IAER,OAAO,QAAQ,CAAC;AAClB,CAAC;AAED;;;;;;GAMG;AACH,MAAM,UAAU,YAAY,CAC1B,QAA4B,EAC5B,KAAa;IAEb,MAAM,UAAU,GAAG,QAAQ,CAAC,KAAK,CAAC,CAAC;IACnC,IAAI,UAAU,CAAC,MAAM,KAAK,CAAC,IAAI,QAAQ,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;QACrD,OAAO,EAAE,CAAC;IACZ,CAAC;IAED,MAAM,IAAI,GAAG,QAAQ,CAAC,GAAG,CAAC,CAAC,OAAO,EAAE,EAAE,CAAC,QAAQ,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC;IAC/D,MAAM,SAAS,GACb,IAAI,CAAC,MAAM,CAAC,CAAC,GAAG,EAAE,GAAG,EAAE,EAAE,CAAC,GAAG,GAAG,GAAG,CAAC,MAAM,EAAE,CAAC,CAAC,GAAG,CAAC,IAAI,CAAC,MAAM,IAAI,CAAC,CAAC,CAAC;IAEtE,mEAAmE;IACnE,MAAM,OAAO,GAAG,IAAI,GAAG,EAAkB,CAAC;IAC1C,KAAK,MAAM,IAAI,IAAI,IAAI,GAAG,CAAC,UAAU,CAAC,EAAE,CAAC;QACvC,IAAI,KAAK,GAAG,CAAC,CAAC;QACd,KAAK,MAAM,GAAG,IAAI,IAAI,EAAE,CAAC;YACvB,IAAI,GAAG,CAAC,QAAQ,CAAC,IAAI,CAAC,EAAE,CAAC;gBACvB,KAAK,IAAI,CAAC,CAAC;YACb,CAAC;QACH,CAAC;QACD,OAAO,CAAC,GAAG,CAAC,IAAI,EAAE,KAAK,CAAC,CAAC;IAC3B,CAAC;IAED,MAAM,MAAM,GAAc,EAAE,CAAC;IAC7B,QAAQ,CAAC,OAAO,CAAC,CAAC,OAAO,EAAE,KAAK,EAAE,EAAE;QAClC,MAAM,GAAG,GAAG,IAAI,CAAC,KAAK,CAAC,IAAI,EAAE,CAAC;QAC9B,IAAI,GAAG,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;YACrB,OAAO;QACT,CAAC;QACD,MAAM,MAAM,GAAG,IAAI,GAAG,EAAkB,CAAC;QACzC,KAAK,MAAM,KAAK,IAAI,GAAG,EAAE,CAAC;YACxB,MAAM,CAAC,GAAG,CAAC,KAAK,EAAE,CAAC,MAAM,CAAC,GAAG,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC;QAClD,CAAC;QAED,IAAI,KAAK,GAAG,CAAC,CAAC;QACd,KAAK,MAAM,IAAI,IAAI,IAAI,GAAG,CAAC,UAAU,CAAC,EAAE,CAAC;YACvC,MAAM,IAAI,GAAG,MAAM,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC;YAC9B,IAAI,IAAI,KAAK,SAAS,EAAE,CAAC;gBACvB,SAAS;YACX,CAAC;YACD,MAAM,CAAC,GAAG,OAAO,CAAC,GAAG,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;YACjC,mEAAmE;YACnE,MAAM,GAAG,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,GAAG,CAAC,IAAI,CAAC,MAAM,GAAG,CAAC,GAAG,GAAG,CAAC,GAAG,CAAC,CAAC,GAAG,GAAG,CAAC,CAAC,CAAC;YAC9D,MAAM,IAAI,GAAG,IAAI,GAAG,CAAC,EAAE,GAAG,CAAC,CAAC,CAAC;YAC7B,MAAM,KAAK,GAAG,IAAI,GAAG,EAAE,GAAG,CAAC,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC,GAAG,GAAG,CAAC,MAAM,CAAC,GAAG,CAAC,SAAS,IAAI,CAAC,CAAC,CAAC,CAAC;YACxE,KAAK,IAAI,GAAG,GAAG,CAAC,IAAI,GAAG,KAAK,CAAC,CAAC;QAChC,CAAC;QAED,IAAI,KAAK,GAAG,CAAC,EAAE,CAAC;YACd,MAAM,CAAC,IAAI,CAAC,EAAE,GAAG,OAAO,EAAE,KAAK,EAAE,CAAC,CAAC;QACrC,CAAC;IACH,CAAC,CAAC,CAAC;IAEH,OAAO,MAAM,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,KAAK,GAAG,CAAC,CAAC,KAAK,CAAC,CAAC;AAClD,CAAC;AAED;;;;;;GAMG;AACH,MAAM,UAAU,cAAc,CAC5B,OAAe,EACf,KAAa,EACb,KAAK,GAAG,CAAC;IAET,OAAO,YAAY,CAAC,aAAa,CAAC,OAAO,CAAC,EAAE,KAAK,CAAC,CAAC,KAAK,CAAC,CAAC,EAAE,KAAK,CAAC,CAAC;AACrE,CAAC"}

package/dist/tools/research.d.ts

@@ -12,6 +12,7 @@ import { z } from "zod";
 import type { McpServer } from "@modelcontextprotocol/sdk/server/mcp.js";
 import { type RoutedPage } from "../extraction/router.js";
 import { Cache } from "../utils/cache.js";
+import { type IndependenceResult } from "../ranking/independence.js";
 import type { SearchSource } from "./search.js";
 /** Input schema for `web_research`. */
 export declare const researchInputSchema: {
@@ -21,6 +22,15 @@ export declare const researchInputSchema: {
     count_per_query: z.ZodDefault<z.ZodNumber>;
 };
 /** The spec-shaped `web_research` response. */
+/** Whether a source is an independent account or a copy of another. */
+export interface SourceIndependence {
+    /** Index of the content cluster this source belongs to. */
+    cluster: number;
+    /** Whether this source is the representative of its cluster. */
+    primary: boolean;
+    /** How many other opened sources share this cluster. */
+    duplicates: number;
+}
 /** A span of a source page that supports the answer. */
 export interface Citation {
     /** 1-based index into `sources`. */
@@ -47,6 +57,34 @@ export interface ResearchResponse {
      * statement to a span rather than to a whole page.
      */
     citations: Citation[];
+    /**
+     * Countable facts about the evidence behind the answer.
+     *
+     * Every field is something the server measured, not a judgement: there is no
+     * language model here to score truth or agreement, and an invented
+     * confidence number would be worse than none. An agent can use these to
+     * decide whether to research further.
+     */
+    evidence: {
+        /** Pages actually opened and extracted. */
+        sources_opened: number;
+        /**
+         * Distinct content clusters among those pages. Syndicated reprints and
+         * multiple pages from one host collapse into a single cluster, so this is
+         * the number of genuinely separate accounts.
+         */
+        independent_sources: number;
+        /** Pages that duplicate another page's content or host. */
+        derivative_sources: number;
+        /**
+         * Fraction of the question's content words that appear in the cited
+         * passages, 0..1. Low values mean the citations may not address the
+         * question, not that the answer is wrong.
+         */
+        query_term_coverage: number;
+        /** Number of cited spans in `citations`. */
+        cited_spans: number;
+    };
     extraction_stats: {
         method_counts: Record<string, number>;
         avgConfidence: number;
@@ -62,7 +100,15 @@ export interface ResearchResponse {
  * best ones returned with their offsets, which raises the signal and cuts the
  * token count, since only matching paragraphs are included.
  */
-export declare function collectCitations(pages: readonly RoutedPage[], sources: readonly SearchSource[], question: string, limit?: number): Citation[];
+export declare function collectCitations(pages: readonly RoutedPage[], sources: readonly SearchSource[], question: string, limit?: number, independence?: readonly IndependenceResult[]): Citation[];
+/**
+ * Fraction of the question's content words that appear in the cited passages.
+ *
+ * A blunt but honest measure of whether the citations engage with the question
+ * at all. It says nothing about correctness, which is not something this
+ * server can determine.
+ */
+export declare function queryTermCoverage(question: string, citations: readonly Citation[]): number;
 /**
  * Run a deep research query and return the spec-shaped response.
  *

package/dist/tools/research.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"research.d.ts","sourceRoot":"","sources":["../../src/tools/research.ts"],"names":[],"mappings":"AAAA;;;;;;;;;GASG;AAEH,OAAO,EAAE,CAAC,EAAE,MAAM,KAAK,CAAC;AACxB,OAAO,KAAK,EAAE,SAAS,EAAE,MAAM,yCAAyC,CAAC;AAOzE,OAAO,EAAW,KAAK,UAAU,EAAE,MAAM,yBAAyB,CAAC;AACnE,OAAO,EAAE,KAAK,EAAE,MAAM,mBAAmB,CAAC;AAE1C,OAAO,KAAK,EAAE,YAAY,EAAE,MAAM,aAAa,CAAC;AAEhD,uCAAuC;AACvC,eAAO,MAAM,mBAAmB;;;;;CAK/B,CAAC;AAEF,+CAA+C;AAC/C,wDAAwD;AACxD,MAAM,WAAW,QAAQ;IACvB,oCAAoC;IACpC,QAAQ,EAAE,MAAM,CAAC;IACjB,8CAA8C;IAC9C,GAAG,EAAE,MAAM,CAAC;IACZ,wBAAwB;IACxB,KAAK,EAAE,MAAM,CAAC;IACd,gEAAgE;IAChE,KAAK,EAAE,MAAM,CAAC;IACd,0EAA0E;IAC1E,KAAK,EAAE,MAAM,CAAC;IACd,yCAAyC;IACzC,GAAG,EAAE,MAAM,CAAC;IACZ,iEAAiE;IACjE,SAAS,EAAE,MAAM,CAAC;CACnB;AAED,MAAM,WAAW,gBAAgB;IAC/B,MAAM,EAAE,MAAM,CAAC;IACf,OAAO,EAAE,YAAY,EAAE,CAAC;IACxB,YAAY,EAAE,MAAM,EAAE,CAAC;IACvB;;;OAGG;IACH,SAAS,EAAE,QAAQ,EAAE,CAAC;IACtB,gBAAgB,EAAE;QAChB,aAAa,EAAE,MAAM,CAAC,MAAM,EAAE,MAAM,CAAC,CAAC;QACtC,aAAa,EAAE,MAAM,CAAC;KACvB,CAAC;CACH;AAmGD,6EAA6E;AAC7E;;;;;;;;GAQG;AACH,wBAAgB,gBAAgB,CAC9B,KAAK,EAAE,SAAS,UAAU,EAAE,EAC5B,OAAO,EAAE,SAAS,YAAY,EAAE,EAChC,QAAQ,EAAE,MAAM,EAChB,KAAK,SAAe,GACnB,QAAQ,EAAE,CAqDZ;AAkCD;;;;;GAKG;AACH,wBAAsB,WAAW,CAAC,IAAI,EAAE;IACtC,QAAQ,EAAE,MAAM,CAAC;IACjB,KAAK,CAAC,EAAE,OAAO,GAAG,MAAM,CAAC;IACzB,YAAY,CAAC,EAAE,MAAM,CAAC;IACtB,eAAe,CAAC,EAAE,MAAM,CAAC;IACzB,KAAK,CAAC,EAAE,KAAK,CAAC;CACf,GAAG,OAAO,CAAC,gBAAgB,CAAC,CAgE5B;AAED,sFAAsF;AACtF,wBAAgB,gBAAgB,CAAC,MAAM,EAAE,SAAS,EAAE,KAAK,CAAC,EAAE,KAAK,GAAG,IAAI,CAgCvE"}
+{"version":3,"file":"research.d.ts","sourceRoot":"","sources":["../../src/tools/research.ts"],"names":[],"mappings":"AAAA;;;;;;;;;GASG;AAEH,OAAO,EAAE,CAAC,EAAE,MAAM,KAAK,CAAC;AACxB,OAAO,KAAK,EAAE,SAAS,EAAE,MAAM,yCAAyC,CAAC;AAOzE,OAAO,EAAW,KAAK,UAAU,EAAE,MAAM,yBAAyB,CAAC;AACnE,OAAO,EAAE,KAAK,EAAE,MAAM,mBAAmB,CAAC;AAE1C,OAAO,EAGL,KAAK,kBAAkB,EACxB,MAAM,4BAA4B,CAAC;AACpC,OAAO,KAAK,EAAE,YAAY,EAAE,MAAM,aAAa,CAAC;AAEhD,uCAAuC;AACvC,eAAO,MAAM,mBAAmB;;;;;CAK/B,CAAC;AAEF,+CAA+C;AAC/C,uEAAuE;AACvE,MAAM,WAAW,kBAAkB;IACjC,2DAA2D;IAC3D,OAAO,EAAE,MAAM,CAAC;IAChB,gEAAgE;IAChE,OAAO,EAAE,OAAO,CAAC;IACjB,wDAAwD;IACxD,UAAU,EAAE,MAAM,CAAC;CACpB;AAED,wDAAwD;AACxD,MAAM,WAAW,QAAQ;IACvB,oCAAoC;IACpC,QAAQ,EAAE,MAAM,CAAC;IACjB,8CAA8C;IAC9C,GAAG,EAAE,MAAM,CAAC;IACZ,wBAAwB;IACxB,KAAK,EAAE,MAAM,CAAC;IACd,gEAAgE;IAChE,KAAK,EAAE,MAAM,CAAC;IACd,0EAA0E;IAC1E,KAAK,EAAE,MAAM,CAAC;IACd,yCAAyC;IACzC,GAAG,EAAE,MAAM,CAAC;IACZ,iEAAiE;IACjE,SAAS,EAAE,MAAM,CAAC;CACnB;AAED,MAAM,WAAW,gBAAgB;IAC/B,MAAM,EAAE,MAAM,CAAC;IACf,OAAO,EAAE,YAAY,EAAE,CAAC;IACxB,YAAY,EAAE,MAAM,EAAE,CAAC;IACvB;;;OAGG;IACH,SAAS,EAAE,QAAQ,EAAE,CAAC;IACtB;;;;;;;OAOG;IACH,QAAQ,EAAE;QACR,2CAA2C;QAC3C,cAAc,EAAE,MAAM,CAAC;QACvB;;;;WAIG;QACH,mBAAmB,EAAE,MAAM,CAAC;QAC5B,2DAA2D;QAC3D,kBAAkB,EAAE,MAAM,CAAC;QAC3B;;;;WAIG;QACH,mBAAmB,EAAE,MAAM,CAAC;QAC5B,4CAA4C;QAC5C,WAAW,EAAE,MAAM,CAAC;KACrB,CAAC;IACF,gBAAgB,EAAE;QAChB,aAAa,EAAE,MAAM,CAAC,MAAM,EAAE,MAAM,CAAC,CAAC;QACtC,aAAa,EAAE,MAAM,CAAC;KACvB,CAAC;CACH;AAmGD,6EAA6E;AAC7E;;;;;;;;GAQG;AACH,wBAAgB,gBAAgB,CAC9B,KAAK,EAAE,SAAS,UAAU,EAAE,EAC5B,OAAO,EAAE,SAAS,YAAY,EAAE,EAChC,QAAQ,EAAE,MAAM,EAChB,KAAK,SAAe,EACpB,YAAY,GAAE,SAAS,kBAAkB,EAAO,GAC/C,QAAQ,EAAE,CAkEZ;AAgBD;;;;;;GAMG;AACH,wBAAgB,iBAAiB,CAC/B,QAAQ,EAAE,MAAM,EAChB,SAAS,EAAE,SAAS,QAAQ,EAAE,GAC7B,MAAM,CAaR;AAoBD;;;;;GAKG;AACH,wBAAsB,WAAW,CAAC,IAAI,EAAE;IACtC,QAAQ,EAAE,MAAM,CAAC;IACjB,KAAK,CAAC,EAAE,OAAO,GAAG,MAAM,CAAC;IACzB,YAAY,CAAC,EAAE,MAAM,CAAC;IACtB,eAAe,CAAC,EAAE,MAAM,CAAC;IACzB,KAAK,CAAC,EAAE,KAAK,CAAC;CACf,GAAG,OAAO,CAAC,gBAAgB,CAAC,CAuF5B;AAED,sFAAsF;AACtF,wBAAgB,gBAAgB,CAAC,MAAM,EAAE,SAAS,EAAE,KAAK,CAAC,EAAE,KAAK,GAAG,IAAI,CAgCvE"}

package/dist/tools/research.js

@@ -16,7 +16,8 @@ import { expandQueries } from "../utils/queries.js";
 import { freshnessFromRecencyDays } from "../providers/brave.js";
 import { getPage } from "../extraction/router.js";
 import { Cache } from "../utils/cache.js";
-import { selectPassages } from "../ranking/passages.js";
+import { selectPassages, tokenize } from "../ranking/passages.js";
+import { analyzeIndependence, countIndependent, } from "../ranking/independence.js";
 /** Input schema for `web_research`. */
 export const researchInputSchema = {
     question: z.string().min(1, "question must not be empty"),
@@ -105,13 +106,19 @@ async function openPages(urls, cache) {
  * best ones returned with their offsets, which raises the signal and cuts the
  * token count, since only matching paragraphs are included.
  */
-export function collectCitations(pages, sources, question, limit = MAX_EXCERPTS) {
+export function collectCitations(pages, sources, question, limit = MAX_EXCERPTS, independence = []) {
     const candidates = [];
     pages.forEach((page, index) => {
         const content = page.content.trim();
         if (content === "") {
             return;
         }
+        // A derivative source republishes another source's text. Quoting it adds
+        // no evidence and, once the breadth pass has run, the remaining slots
+        // would otherwise be filled with the same story under a second byline.
+        if (independence[index]?.primary === false) {
+            return;
+        }
         const source = sources[index];
         for (const passage of selectPassages(content, question, PASSAGES_PER_PAGE)) {
             candidates.push({
@@ -127,19 +134,24 @@ export function collectCitations(pages, sources, question, limit = MAX_EXCERPTS)
         }
     });
     candidates.sort((a, b) => b.score - a.score);
-    // Prefer breadth: take the best passage from each distinct source before a
-    // second from any one of them, so the answer is not a single page's view.
+    // Prefer breadth across independent accounts. Spreading over distinct URLs
+    // is not enough: five outlets running one wire story would fill every slot
+    // with the same text and present it as five corroborating sources.
+    const groupOf = (citation) => independence[citation - 1]?.cluster ?? citation;
     const chosen = [];
-    const used = new Set();
+    const usedGroups = new Set();
     for (const candidate of candidates) {
         if (chosen.length >= limit) {
             break;
         }
-        if (!used.has(candidate.citation)) {
-            used.add(candidate.citation);
+        const group = groupOf(candidate.citation);
+        if (!usedGroups.has(group)) {
+            usedGroups.add(group);
             chosen.push(candidate);
         }
     }
+    // Remaining slots go to further passages from those same independent
+    // accounts, which adds depth without re-reporting one story twice.
     for (const candidate of candidates) {
         if (chosen.length >= limit) {
             break;
@@ -165,6 +177,27 @@ function buildAnswer(pages, citations) {
         .join("\n\n");
     return `Based on ${pages.length} sources:\n\n${body}`;
 }
+/**
+ * Fraction of the question's content words that appear in the cited passages.
+ *
+ * A blunt but honest measure of whether the citations engage with the question
+ * at all. It says nothing about correctness, which is not something this
+ * server can determine.
+ */
+export function queryTermCoverage(question, citations) {
+    const terms = new Set(tokenize(question));
+    if (terms.size === 0) {
+        return 0;
+    }
+    const cited = new Set(tokenize(citations.map((c) => c.quote).join(" ")));
+    let hits = 0;
+    for (const term of terms) {
+        if (cited.has(term)) {
+            hits += 1;
+        }
+    }
+    return Number((hits / terms.size).toFixed(4));
+}
 /** Aggregate extraction stats across the opened pages. */
 function extractionStats(pages) {
     const methodCounts = {};
@@ -219,13 +252,27 @@ export async function runResearch(args) {
     const sources = top.map(toSource);
     // 4) Open the top pages in parallel (bounded concurrency, per-page timeout).
     const pages = await openPages(top.map((r) => r.url), cache);
-    // 5) Select the spans that address the question, and cite them.
-    const citations = collectCitations(pages, sources, question);
+    // 5) Work out which pages are genuinely separate accounts before citing, so
+    // syndicated copies do not masquerade as corroboration.
+    const independence = analyzeIndependence(pages.map((page, index) => ({
+        url: sources[index]?.url ?? page.url,
+        content: page.content,
+    })));
+    // 6) Select the spans that address the question, and cite them.
+    const citations = collectCitations(pages, sources, question, MAX_EXCERPTS, independence);
+    const independentSources = countIndependent(independence);
     const response = {
         answer: buildAnswer(pages, citations),
         sources,
         queries_used: queries,
         citations,
+        evidence: {
+            sources_opened: pages.length,
+            independent_sources: independentSources,
+            derivative_sources: Math.max(0, pages.length - independentSources),
+            query_term_coverage: queryTermCoverage(question, citations),
+            cited_spans: citations.length,
+        },
         extraction_stats: extractionStats(pages),
     };
     cache?.setSearch(cacheKey, question, response);

package/dist/tools/research.js.map

@@ -1 +1 @@
-{"version":3,"file":"research.js","sourceRoot":"","sources":["../../src/tools/research.ts"],"names":[],"mappings":"AAAA;;;;;;;;;GASG;AAEH,OAAO,EAAE,CAAC,EAAE,MAAM,KAAK,CAAC;AAGxB,OAAO,EAAE,eAAe,EAAE,MAAM,uBAAuB,CAAC;AACxD,OAAO,EAAE,WAAW,EAAE,MAAM,2BAA2B,CAAC;AACxD,OAAO,EAAE,MAAM,EAAqB,MAAM,sBAAsB,CAAC;AACjE,OAAO,EAAE,aAAa,EAAE,MAAM,qBAAqB,CAAC;AACpD,OAAO,EAAE,wBAAwB,EAAE,MAAM,uBAAuB,CAAC;AACjE,OAAO,EAAE,OAAO,EAAmB,MAAM,yBAAyB,CAAC;AACnE,OAAO,EAAE,KAAK,EAAE,MAAM,mBAAmB,CAAC;AAC1C,OAAO,EAAE,cAAc,EAAE,MAAM,wBAAwB,CAAC;AAGxD,uCAAuC;AACvC,MAAM,CAAC,MAAM,mBAAmB,GAAG;IACjC,QAAQ,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,GAAG,CAAC,CAAC,EAAE,4BAA4B,CAAC;IACzD,KAAK,EAAE,CAAC,CAAC,IAAI,CAAC,CAAC,OAAO,EAAE,MAAM,CAAC,CAAC,CAAC,OAAO,CAAC,MAAM,CAAC;IAChD,YAAY,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,GAAG,EAAE,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC,QAAQ,EAAE;IACzD,eAAe,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,GAAG,EAAE,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,EAAE,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC;CAC5D,CAAC;AAoCF,4BAA4B;AAC5B,MAAM,UAAU,GAAG,UAAU,CAAC;AAE9B,oDAAoD;AACpD,MAAM,eAAe,GAAG,KAAK,CAAC;AAE9B,2CAA2C;AAC3C,MAAM,WAAW,GAAG,CAAC,CAAC;AAEtB,sDAAsD;AACtD,MAAM,SAAS,GAAG,EAAE,CAAC;AAErB,wEAAwE;AACxE,MAAM,iBAAiB,GAAG,CAAC,CAAC;AAE5B,yDAAyD;AACzD,MAAM,YAAY,GAAG,CAAC,CAAC;AAEvB,2DAA2D;AAC3D,SAAS,QAAQ,CAAC,MAAoB;IACpC,OAAO;QACL,KAAK,EAAE,MAAM,CAAC,KAAK;QACnB,GAAG,EAAE,MAAM,CAAC,GAAG;QACf,OAAO,EAAE,MAAM,CAAC,OAAO;QACvB,SAAS,EAAE,MAAM,CAAC,SAAS;QAC3B,SAAS,EAAE,MAAM,CAAC,SAAS;KAC5B,CAAC;AACJ,CAAC;AAED,gEAAgE;AAChE,SAAS,YAAY,CAAC,SAA6B,EAAE,WAAmB;IACtE,IAAI,SAAS,KAAK,SAAS,IAAI,SAAS,KAAK,EAAE,EAAE,CAAC;QAChD,OAAO,IAAI,CAAC;IACd,CAAC;IACD,MAAM,IAAI,GAAG,IAAI,CAAC,KAAK,CAAC,SAAS,CAAC,CAAC;IACnC,IAAI,MAAM,CAAC,KAAK,CAAC,IAAI,CAAC,EAAE,CAAC;QACvB,OAAO,IAAI,CAAC;IACd,CAAC;IACD,OAAO,IAAI,IAAI,IAAI,CAAC,GAAG,EAAE,GAAG,WAAW,GAAG,UAAU,CAAC;AACvD,CAAC;AAOD,4DAA4D;AAC5D,KAAK,UAAU,QAAQ,CAAC,GAAW,EAAE,KAAwB;IAC3D,IAAI,KAAgD,CAAC;IACrD,MAAM,OAAO,GAAG,IAAI,OAAO,CAAQ,CAAC,CAAC,EAAE,MAAM,EAAE,EAAE;QAC/C,KAAK,GAAG,UAAU,CAChB,GAAG,EAAE,CAAC,MAAM,CAAC,IAAI,KAAK,CAAC,mCAAmC,eAAe,IAAI,CAAC,CAAC,EAC/E,eAAe,CAChB,CAAC;IACJ,CAAC,CAAC,CAAC;IAEH,IAAI,CAAC;QACH,MAAM,IAAI,GAAG,MAAM,OAAO,CAAC,IAAI,CAAC,CAAC,OAAO,CAAC,GAAG,EAAE,EAAE,EAAE,EAAE,KAAK,EAAE,CAAC,EAAE,OAAO,CAAC,CAAC,CAAC;QACxE,OAAO,EAAE,EAAE,EAAE,IAAI,EAAE,IAAI,EAAE,CAAC;IAC5B,CAAC;IAAC,OAAO,KAAK,EAAE,CAAC;QACf,MAAM,MAAM,GAAG,KAAK,YAAY,KAAK,CAAC,CAAC,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC;QACtE,OAAO,EAAE,EAAE,EAAE,KAAK,EAAE,MAAM,EAAE,CAAC;IAC/B,CAAC;YAAS,CAAC;QACT,IAAI,KAAK,KAAK,SAAS,EAAE,CAAC;YACxB,YAAY,CAAC,KAAK,CAAC,CAAC;QACtB,CAAC;IACH,CAAC;AACH,CAAC;AAED,mFAAmF;AACnF,KAAK,UAAU,SAAS,CACtB,IAAuB,EACvB,KAAwB;IAExB,MAAM,QAAQ,GAAkB,IAAI,KAAK,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;IACvD,IAAI,MAAM,GAAG,CAAC,CAAC;IAEf,KAAK,UAAU,MAAM;QACnB,OAAO,IAAI,EAAE,CAAC;YACZ,MAAM,KAAK,GAAG,MAAM,CAAC;YACrB,MAAM,IAAI,CAAC,CAAC;YACZ,IAAI,KAAK,IAAI,IAAI,CAAC,MAAM,EAAE,CAAC;gBACzB,OAAO;YACT,CAAC;YACD,QAAQ,CAAC,KAAK,CAAC,GAAG,MAAM,QAAQ,CAAC,IAAI,CAAC,KAAK,CAAC,IAAI,EAAE,EAAE,KAAK,CAAC,CAAC;QAC7D,CAAC;IACH,CAAC;IAED,MAAM,OAAO,GAAG,KAAK,CAAC,IAAI,CACxB,EAAE,MAAM,EAAE,IAAI,CAAC,GAAG,CAAC,WAAW,EAAE,IAAI,CAAC,MAAM,CAAC,EAAE,EAC9C,GAAG,EAAE,CAAC,MAAM,EAAE,CACf,CAAC;IACF,MAAM,OAAO,CAAC,GAAG,CAAC,OAAO,CAAC,CAAC;IAE3B,OAAO,QAAQ,CAAC,OAAO,CAAC,CAAC,OAAO,EAAE,EAAE,CAAC,CAAC,OAAO,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC;AAC3E,CAAC;AAED,6EAA6E;AAC7E;;;;;;;;GAQG;AACH,MAAM,UAAU,gBAAgB,CAC9B,KAA4B,EAC5B,OAAgC,EAChC,QAAgB,EAChB,KAAK,GAAG,YAAY;IAEpB,MAAM,UAAU,GAAwC,EAAE,CAAC;IAE3D,KAAK,CAAC,OAAO,CAAC,CAAC,IAAI,EAAE,KAAK,EAAE,EAAE;QAC5B,MAAM,OAAO,GAAG,IAAI,CAAC,OAAO,CAAC,IAAI,EAAE,CAAC;QACpC,IAAI,OAAO,KAAK,EAAE,EAAE,CAAC;YACnB,OAAO;QACT,CAAC;QACD,MAAM,MAAM,GAAG,OAAO,CAAC,KAAK,CAAC,CAAC;QAC9B,KAAK,MAAM,OAAO,IAAI,cAAc,CAAC,OAAO,EAAE,QAAQ,EAAE,iBAAiB,CAAC,EAAE,CAAC;YAC3E,UAAU,CAAC,IAAI,CAAC;gBACd,QAAQ,EAAE,KAAK,GAAG,CAAC;gBACnB,GAAG,EAAE,MAAM,EAAE,GAAG,IAAI,IAAI,CAAC,GAAG;gBAC5B,KAAK,EAAE,MAAM,EAAE,KAAK,IAAI,IAAI,CAAC,KAAK;gBAClC,KAAK,EAAE,OAAO,CAAC,IAAI;gBACnB,KAAK,EAAE,OAAO,CAAC,KAAK;gBACpB,GAAG,EAAE,OAAO,CAAC,GAAG;gBAChB,SAAS,EAAE,MAAM,CAAC,OAAO,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC;gBAC3C,KAAK,EAAE,OAAO,CAAC,KAAK;aACrB,CAAC,CAAC;QACL,CAAC;IACH,CAAC,CAAC,CAAC;IAEH,UAAU,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,KAAK,GAAG,CAAC,CAAC,KAAK,CAAC,CAAC;IAE7C,2EAA2E;IAC3E,0EAA0E;IAC1E,MAAM,MAAM,GAAwC,EAAE,CAAC;IACvD,MAAM,IAAI,GAAG,IAAI,GAAG,EAAU,CAAC;IAC/B,KAAK,MAAM,SAAS,IAAI,UAAU,EAAE,CAAC;QACnC,IAAI,MAAM,CAAC,MAAM,IAAI,KAAK,EAAE,CAAC;YAC3B,MAAM;QACR,CAAC;QACD,IAAI,CAAC,IAAI,CAAC,GAAG,CAAC,SAAS,CAAC,QAAQ,CAAC,EAAE,CAAC;YAClC,IAAI,CAAC,GAAG,CAAC,SAAS,CAAC,QAAQ,CAAC,CAAC;YAC7B,MAAM,CAAC,IAAI,CAAC,SAAS,CAAC,CAAC;QACzB,CAAC;IACH,CAAC;IACD,KAAK,MAAM,SAAS,IAAI,UAAU,EAAE,CAAC;QACnC,IAAI,MAAM,CAAC,MAAM,IAAI,KAAK,EAAE,CAAC;YAC3B,MAAM;QACR,CAAC;QACD,IAAI,CAAC,MAAM,CAAC,QAAQ,CAAC,SAAS,CAAC,EAAE,CAAC;YAChC,MAAM,CAAC,IAAI,CAAC,SAAS,CAAC,CAAC;QACzB,CAAC;IACH,CAAC;IAED,OAAO,MAAM;SACV,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,QAAQ,GAAG,CAAC,CAAC,QAAQ,IAAI,CAAC,CAAC,KAAK,GAAG,CAAC,CAAC,KAAK,CAAC;SAC5D,GAAG,CAAC,CAAC,EAAE,KAAK,EAAE,GAAG,QAAQ,EAAE,EAAE,EAAE;QAC9B,KAAK,KAAK,CAAC;QACX,OAAO,QAAQ,CAAC;IAClB,CAAC,CAAC,CAAC;AACP,CAAC;AAED,iEAAiE;AACjE,SAAS,WAAW,CAClB,KAA4B,EAC5B,SAA8B;IAE9B,IAAI,KAAK,CAAC,MAAM,KAAK,CAAC,IAAI,SAAS,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;QACjD,OAAO,kBAAkB,CAAC;IAC5B,CAAC;IACD,MAAM,IAAI,GAAG,SAAS;SACnB,GAAG,CAAC,CAAC,QAAQ,EAAE,EAAE,CAAC,IAAI,QAAQ,CAAC,QAAQ,KAAK,QAAQ,CAAC,KAAK,EAAE,CAAC;SAC7D,IAAI,CAAC,MAAM,CAAC,CAAC;IAChB,OAAO,YAAY,KAAK,CAAC,MAAM,gBAAgB,IAAI,EAAE,CAAC;AACxD,CAAC;AAED,0DAA0D;AAC1D,SAAS,eAAe,CAAC,KAA4B;IAInD,MAAM,YAAY,GAA2B,EAAE,CAAC;IAChD,IAAI,aAAa,GAAG,CAAC,CAAC;IAEtB,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;QACzB,MAAM,MAAM,GAAG,IAAI,CAAC,UAAU,CAAC,MAAM,CAAC;QACtC,YAAY,CAAC,MAAM,CAAC,GAAG,CAAC,YAAY,CAAC,MAAM,CAAC,IAAI,CAAC,CAAC,GAAG,CAAC,CAAC;QACvD,aAAa,IAAI,IAAI,CAAC,UAAU,CAAC,UAAU,CAAC;IAC9C,CAAC;IAED,MAAM,aAAa,GAAG,KAAK,CAAC,MAAM,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,aAAa,GAAG,KAAK,CAAC,MAAM,CAAC;IAC5E,OAAO,EAAE,aAAa,EAAE,YAAY,EAAE,aAAa,EAAE,CAAC;AACxD,CAAC;AAED;;;;;GAKG;AACH,MAAM,CAAC,KAAK,UAAU,WAAW,CAAC,IAMjC;IACC,MAAM,QAAQ,GAAG,IAAI,CAAC,QAAQ,CAAC;IAC/B,MAAM,KAAK,GAAG,IAAI,CAAC,KAAK,IAAI,MAAM,CAAC;IACnC,MAAM,WAAW,GAAG,IAAI,CAAC,YAAY,CAAC;IACtC,MAAM,aAAa,GAAG,IAAI,CAAC,eAAe,IAAI,CAAC,CAAC;IAChD,MAAM,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC;IAEzB,MAAM,QAAQ,GAAG,YAAY,QAAQ,IAAI,KAAK,IAAI,WAAW,IAAI,KAAK,IAAI,aAAa,EAAE,CAAC;IAC1F,MAAM,MAAM,GAAG,KAAK,EAAE,SAAS,CAAC,QAAQ,CAAC,CAAC;IAC1C,IAAI,MAAM,KAAK,IAAI,IAAI,MAAM,KAAK,SAAS,EAAE,CAAC;QAC5C,OAAO,MAA0B,CAAC;IACpC,CAAC;IAED,kEAAkE;IAClE,MAAM,QAAQ,GAAG,aAAa,CAAC,QAAQ,CAAC,CAAC;IACzC,MAAM,OAAO,GAAG,KAAK,KAAK,OAAO,CAAC,CAAC,CAAC,QAAQ,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC,QAAQ,CAAC;IAEpE,oDAAoD;IACpD,MAAM,SAAS,GAAG,WAAW,KAAK,SAAS;QACzC,CAAC,CAAC,SAAS;QACX,CAAC,CAAC,wBAAwB,CAAC,WAAW,CAAC,CAAC;IAC1C,MAAM,OAAO,GAAG,MAAM,OAAO,CAAC,GAAG,CAC/B,OAAO,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CACpB,eAAe,CAAC,KAAK,EAAE;QACrB,KAAK,EAAE,aAAa;QACpB,SAAS;QACT,YAAY,EAAE,WAAW;QACzB,aAAa,EAAE,IAAI;KACpB,CAAC,CACH,CACF,CAAC;IAEF,yEAAyE;IACzE,MAAM,MAAM,GAAG,OAAO,CAAC,IAAI,EAAE,CAAC;IAC9B,MAAM,OAAO,GAAG,WAAW,CAAC,MAAM,CAAC,CAAC;IACpC,MAAM,MAAM,GAAG,MAAM,CAAC,OAAO,EAAE,QAAQ,EAAE,WAAW,IAAI,EAAE,CAAC,CAAC;IAE5D,iDAAiD;IACjD,MAAM,QAAQ,GAAG,WAAW,KAAK,SAAS;QACxC,CAAC,CAAC,MAAM;QACR,CAAC,CAAC,MAAM,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,YAAY,CAAC,CAAC,CAAC,SAAS,EAAE,WAAW,CAAC,CAAC,CAAC;IAEjE,MAAM,GAAG,GAAG,QAAQ,CAAC,KAAK,CAAC,CAAC,EAAE,SAAS,CAAC,CAAC;IACzC,MAAM,OAAO,GAAG,GAAG,CAAC,GAAG,CAAC,QAAQ,CAAC,CAAC;IAElC,6EAA6E;IAC7E,MAAM,KAAK,GAAG,MAAM,SAAS,CAC3B,GAAG,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,GAAG,CAAC,EACrB,KAAK,CACN,CAAC;IAEF,gEAAgE;IAChE,MAAM,SAAS,GAAG,gBAAgB,CAAC,KAAK,EAAE,OAAO,EAAE,QAAQ,CAAC,CAAC;IAE7D,MAAM,QAAQ,GAAqB;QACjC,MAAM,EAAE,WAAW,CAAC,KAAK,EAAE,SAAS,CAAC;QACrC,OAAO;QACP,YAAY,EAAE,OAAO;QACrB,SAAS;QACT,gBAAgB,EAAE,eAAe,CAAC,KAAK,CAAC;KACzC,CAAC;IAEF,KAAK,EAAE,SAAS,CAAC,QAAQ,EAAE,QAAQ,EAAE,QAAQ,CAAC,CAAC;IAC/C,OAAO,QAAQ,CAAC;AAClB,CAAC;AAED,sFAAsF;AACtF,MAAM,UAAU,gBAAgB,CAAC,MAAiB,EAAE,KAAa;IAC/D,MAAM,CAAC,YAAY,CACjB,cAAc,EACd;QACE,KAAK,EAAE,cAAc;QACrB,WAAW,EACT,gTAAgT;QAClT,WAAW,EAAE,mBAAmB;KACjC,EACD,KAAK,EAAE,IAAI,EAAE,EAAE;QACb,MAAM,QAAQ,GAAG,MAAM,WAAW,CAAC,EAAE,GAAG,IAAI,EAAE,KAAK,EAAE,CAAC,CAAC;QACvD,OAAO;YACL,OAAO,EAAE,CAAC,EAAE,IAAI,EAAE,MAAM,EAAE,IAAI,EAAE,IAAI,CAAC,SAAS,CAAC,QAAQ,CAAC,EAAE,CAAC;SAC5D,CAAC;IACJ,CAAC,CACF,CAAC;IAEF,MAAM,CAAC,YAAY,CACjB,aAAa,EACb;QACE,KAAK,EAAE,aAAa;QACpB,WAAW,EACT,uNAAuN;QACzN,WAAW,EAAE,mBAAmB;KACjC,EACD,KAAK,EAAE,IAAI,EAAE,EAAE;QACb,MAAM,QAAQ,GAAG,MAAM,WAAW,CAAC,EAAE,GAAG,IAAI,EAAE,KAAK,EAAE,CAAC,CAAC;QACvD,OAAO;YACL,OAAO,EAAE,CAAC,EAAE,IAAI,EAAE,MAAM,EAAE,IAAI,EAAE,IAAI,CAAC,SAAS,CAAC,QAAQ,CAAC,EAAE,CAAC;SAC5D,CAAC;IACJ,CAAC,CACF,CAAC;AACJ,CAAC"}
+{"version":3,"file":"research.js","sourceRoot":"","sources":["../../src/tools/research.ts"],"names":[],"mappings":"AAAA;;;;;;;;;GASG;AAEH,OAAO,EAAE,CAAC,EAAE,MAAM,KAAK,CAAC;AAGxB,OAAO,EAAE,eAAe,EAAE,MAAM,uBAAuB,CAAC;AACxD,OAAO,EAAE,WAAW,EAAE,MAAM,2BAA2B,CAAC;AACxD,OAAO,EAAE,MAAM,EAAqB,MAAM,sBAAsB,CAAC;AACjE,OAAO,EAAE,aAAa,EAAE,MAAM,qBAAqB,CAAC;AACpD,OAAO,EAAE,wBAAwB,EAAE,MAAM,uBAAuB,CAAC;AACjE,OAAO,EAAE,OAAO,EAAmB,MAAM,yBAAyB,CAAC;AACnE,OAAO,EAAE,KAAK,EAAE,MAAM,mBAAmB,CAAC;AAC1C,OAAO,EAAE,cAAc,EAAE,QAAQ,EAAE,MAAM,wBAAwB,CAAC;AAClE,OAAO,EACL,mBAAmB,EACnB,gBAAgB,GAEjB,MAAM,4BAA4B,CAAC;AAGpC,uCAAuC;AACvC,MAAM,CAAC,MAAM,mBAAmB,GAAG;IACjC,QAAQ,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,GAAG,CAAC,CAAC,EAAE,4BAA4B,CAAC;IACzD,KAAK,EAAE,CAAC,CAAC,IAAI,CAAC,CAAC,OAAO,EAAE,MAAM,CAAC,CAAC,CAAC,OAAO,CAAC,MAAM,CAAC;IAChD,YAAY,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,GAAG,EAAE,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC,QAAQ,EAAE;IACzD,eAAe,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,GAAG,EAAE,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,EAAE,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC;CAC5D,CAAC;AA0EF,4BAA4B;AAC5B,MAAM,UAAU,GAAG,UAAU,CAAC;AAE9B,oDAAoD;AACpD,MAAM,eAAe,GAAG,KAAK,CAAC;AAE9B,2CAA2C;AAC3C,MAAM,WAAW,GAAG,CAAC,CAAC;AAEtB,sDAAsD;AACtD,MAAM,SAAS,GAAG,EAAE,CAAC;AAErB,wEAAwE;AACxE,MAAM,iBAAiB,GAAG,CAAC,CAAC;AAE5B,yDAAyD;AACzD,MAAM,YAAY,GAAG,CAAC,CAAC;AAEvB,2DAA2D;AAC3D,SAAS,QAAQ,CAAC,MAAoB;IACpC,OAAO;QACL,KAAK,EAAE,MAAM,CAAC,KAAK;QACnB,GAAG,EAAE,MAAM,CAAC,GAAG;QACf,OAAO,EAAE,MAAM,CAAC,OAAO;QACvB,SAAS,EAAE,MAAM,CAAC,SAAS;QAC3B,SAAS,EAAE,MAAM,CAAC,SAAS;KAC5B,CAAC;AACJ,CAAC;AAED,gEAAgE;AAChE,SAAS,YAAY,CAAC,SAA6B,EAAE,WAAmB;IACtE,IAAI,SAAS,KAAK,SAAS,IAAI,SAAS,KAAK,EAAE,EAAE,CAAC;QAChD,OAAO,IAAI,CAAC;IACd,CAAC;IACD,MAAM,IAAI,GAAG,IAAI,CAAC,KAAK,CAAC,SAAS,CAAC,CAAC;IACnC,IAAI,MAAM,CAAC,KAAK,CAAC,IAAI,CAAC,EAAE,CAAC;QACvB,OAAO,IAAI,CAAC;IACd,CAAC;IACD,OAAO,IAAI,IAAI,IAAI,CAAC,GAAG,EAAE,GAAG,WAAW,GAAG,UAAU,CAAC;AACvD,CAAC;AAOD,4DAA4D;AAC5D,KAAK,UAAU,QAAQ,CAAC,GAAW,EAAE,KAAwB;IAC3D,IAAI,KAAgD,CAAC;IACrD,MAAM,OAAO,GAAG,IAAI,OAAO,CAAQ,CAAC,CAAC,EAAE,MAAM,EAAE,EAAE;QAC/C,KAAK,GAAG,UAAU,CAChB,GAAG,EAAE,CAAC,MAAM,CAAC,IAAI,KAAK,CAAC,mCAAmC,eAAe,IAAI,CAAC,CAAC,EAC/E,eAAe,CAChB,CAAC;IACJ,CAAC,CAAC,CAAC;IAEH,IAAI,CAAC;QACH,MAAM,IAAI,GAAG,MAAM,OAAO,CAAC,IAAI,CAAC,CAAC,OAAO,CAAC,GAAG,EAAE,EAAE,EAAE,EAAE,KAAK,EAAE,CAAC,EAAE,OAAO,CAAC,CAAC,CAAC;QACxE,OAAO,EAAE,EAAE,EAAE,IAAI,EAAE,IAAI,EAAE,CAAC;IAC5B,CAAC;IAAC,OAAO,KAAK,EAAE,CAAC;QACf,MAAM,MAAM,GAAG,KAAK,YAAY,KAAK,CAAC,CAAC,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC;QACtE,OAAO,EAAE,EAAE,EAAE,KAAK,EAAE,MAAM,EAAE,CAAC;IAC/B,CAAC;YAAS,CAAC;QACT,IAAI,KAAK,KAAK,SAAS,EAAE,CAAC;YACxB,YAAY,CAAC,KAAK,CAAC,CAAC;QACtB,CAAC;IACH,CAAC;AACH,CAAC;AAED,mFAAmF;AACnF,KAAK,UAAU,SAAS,CACtB,IAAuB,EACvB,KAAwB;IAExB,MAAM,QAAQ,GAAkB,IAAI,KAAK,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;IACvD,IAAI,MAAM,GAAG,CAAC,CAAC;IAEf,KAAK,UAAU,MAAM;QACnB,OAAO,IAAI,EAAE,CAAC;YACZ,MAAM,KAAK,GAAG,MAAM,CAAC;YACrB,MAAM,IAAI,CAAC,CAAC;YACZ,IAAI,KAAK,IAAI,IAAI,CAAC,MAAM,EAAE,CAAC;gBACzB,OAAO;YACT,CAAC;YACD,QAAQ,CAAC,KAAK,CAAC,GAAG,MAAM,QAAQ,CAAC,IAAI,CAAC,KAAK,CAAC,IAAI,EAAE,EAAE,KAAK,CAAC,CAAC;QAC7D,CAAC;IACH,CAAC;IAED,MAAM,OAAO,GAAG,KAAK,CAAC,IAAI,CACxB,EAAE,MAAM,EAAE,IAAI,CAAC,GAAG,CAAC,WAAW,EAAE,IAAI,CAAC,MAAM,CAAC,EAAE,EAC9C,GAAG,EAAE,CAAC,MAAM,EAAE,CACf,CAAC;IACF,MAAM,OAAO,CAAC,GAAG,CAAC,OAAO,CAAC,CAAC;IAE3B,OAAO,QAAQ,CAAC,OAAO,CAAC,CAAC,OAAO,EAAE,EAAE,CAAC,CAAC,OAAO,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC;AAC3E,CAAC;AAED,6EAA6E;AAC7E;;;;;;;;GAQG;AACH,MAAM,UAAU,gBAAgB,CAC9B,KAA4B,EAC5B,OAAgC,EAChC,QAAgB,EAChB,KAAK,GAAG,YAAY,EACpB,eAA8C,EAAE;IAEhD,MAAM,UAAU,GAAwC,EAAE,CAAC;IAE3D,KAAK,CAAC,OAAO,CAAC,CAAC,IAAI,EAAE,KAAK,EAAE,EAAE;QAC5B,MAAM,OAAO,GAAG,IAAI,CAAC,OAAO,CAAC,IAAI,EAAE,CAAC;QACpC,IAAI,OAAO,KAAK,EAAE,EAAE,CAAC;YACnB,OAAO;QACT,CAAC;QACD,yEAAyE;QACzE,sEAAsE;QACtE,uEAAuE;QACvE,IAAI,YAAY,CAAC,KAAK,CAAC,EAAE,OAAO,KAAK,KAAK,EAAE,CAAC;YAC3C,OAAO;QACT,CAAC;QACD,MAAM,MAAM,GAAG,OAAO,CAAC,KAAK,CAAC,CAAC;QAC9B,KAAK,MAAM,OAAO,IAAI,cAAc,CAAC,OAAO,EAAE,QAAQ,EAAE,iBAAiB,CAAC,EAAE,CAAC;YAC3E,UAAU,CAAC,IAAI,CAAC;gBACd,QAAQ,EAAE,KAAK,GAAG,CAAC;gBACnB,GAAG,EAAE,MAAM,EAAE,GAAG,IAAI,IAAI,CAAC,GAAG;gBAC5B,KAAK,EAAE,MAAM,EAAE,KAAK,IAAI,IAAI,CAAC,KAAK;gBAClC,KAAK,EAAE,OAAO,CAAC,IAAI;gBACnB,KAAK,EAAE,OAAO,CAAC,KAAK;gBACpB,GAAG,EAAE,OAAO,CAAC,GAAG;gBAChB,SAAS,EAAE,MAAM,CAAC,OAAO,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC;gBAC3C,KAAK,EAAE,OAAO,CAAC,KAAK;aACrB,CAAC,CAAC;QACL,CAAC;IACH,CAAC,CAAC,CAAC;IAEH,UAAU,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,KAAK,GAAG,CAAC,CAAC,KAAK,CAAC,CAAC;IAE7C,2EAA2E;IAC3E,2EAA2E;IAC3E,mEAAmE;IACnE,MAAM,OAAO,GAAG,CAAC,QAAgB,EAAU,EAAE,CAC3C,YAAY,CAAC,QAAQ,GAAG,CAAC,CAAC,EAAE,OAAO,IAAI,QAAQ,CAAC;IAElD,MAAM,MAAM,GAAwC,EAAE,CAAC;IACvD,MAAM,UAAU,GAAG,IAAI,GAAG,EAAU,CAAC;IACrC,KAAK,MAAM,SAAS,IAAI,UAAU,EAAE,CAAC;QACnC,IAAI,MAAM,CAAC,MAAM,IAAI,KAAK,EAAE,CAAC;YAC3B,MAAM;QACR,CAAC;QACD,MAAM,KAAK,GAAG,OAAO,CAAC,SAAS,CAAC,QAAQ,CAAC,CAAC;QAC1C,IAAI,CAAC,UAAU,CAAC,GAAG,CAAC,KAAK,CAAC,EAAE,CAAC;YAC3B,UAAU,CAAC,GAAG,CAAC,KAAK,CAAC,CAAC;YACtB,MAAM,CAAC,IAAI,CAAC,SAAS,CAAC,CAAC;QACzB,CAAC;IACH,CAAC;IACD,qEAAqE;IACrE,mEAAmE;IACnE,KAAK,MAAM,SAAS,IAAI,UAAU,EAAE,CAAC;QACnC,IAAI,MAAM,CAAC,MAAM,IAAI,KAAK,EAAE,CAAC;YAC3B,MAAM;QACR,CAAC;QACD,IAAI,CAAC,MAAM,CAAC,QAAQ,CAAC,SAAS,CAAC,EAAE,CAAC;YAChC,MAAM,CAAC,IAAI,CAAC,SAAS,CAAC,CAAC;QACzB,CAAC;IACH,CAAC;IAED,OAAO,MAAM;SACV,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,QAAQ,GAAG,CAAC,CAAC,QAAQ,IAAI,CAAC,CAAC,KAAK,GAAG,CAAC,CAAC,KAAK,CAAC;SAC5D,GAAG,CAAC,CAAC,EAAE,KAAK,EAAE,GAAG,QAAQ,EAAE,EAAE,EAAE;QAC9B,KAAK,KAAK,CAAC;QACX,OAAO,QAAQ,CAAC;IAClB,CAAC,CAAC,CAAC;AACP,CAAC;AAED,iEAAiE;AACjE,SAAS,WAAW,CAClB,KAA4B,EAC5B,SAA8B;IAE9B,IAAI,KAAK,CAAC,MAAM,KAAK,CAAC,IAAI,SAAS,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;QACjD,OAAO,kBAAkB,CAAC;IAC5B,CAAC;IACD,MAAM,IAAI,GAAG,SAAS;SACnB,GAAG,CAAC,CAAC,QAAQ,EAAE,EAAE,CAAC,IAAI,QAAQ,CAAC,QAAQ,KAAK,QAAQ,CAAC,KAAK,EAAE,CAAC;SAC7D,IAAI,CAAC,MAAM,CAAC,CAAC;IAChB,OAAO,YAAY,KAAK,CAAC,MAAM,gBAAgB,IAAI,EAAE,CAAC;AACxD,CAAC;AAED;;;;;;GAMG;AACH,MAAM,UAAU,iBAAiB,CAC/B,QAAgB,EAChB,SAA8B;IAE9B,MAAM,KAAK,GAAG,IAAI,GAAG,CAAC,QAAQ,CAAC,QAAQ,CAAC,CAAC,CAAC;IAC1C,IAAI,KAAK,CAAC,IAAI,KAAK,CAAC,EAAE,CAAC;QACrB,OAAO,CAAC,CAAC;IACX,CAAC;IACD,MAAM,KAAK,GAAG,IAAI,GAAG,CAAC,QAAQ,CAAC,SAAS,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC;IACzE,IAAI,IAAI,GAAG,CAAC,CAAC;IACb,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;QACzB,IAAI,KAAK,CAAC,GAAG,CAAC,IAAI,CAAC,EAAE,CAAC;YACpB,IAAI,IAAI,CAAC,CAAC;QACZ,CAAC;IACH,CAAC;IACD,OAAO,MAAM,CAAC,CAAC,IAAI,GAAG,KAAK,CAAC,IAAI,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC,CAAC;AAChD,CAAC;AAED,0DAA0D;AAC1D,SAAS,eAAe,CAAC,KAA4B;IAInD,MAAM,YAAY,GAA2B,EAAE,CAAC;IAChD,IAAI,aAAa,GAAG,CAAC,CAAC;IAEtB,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;QACzB,MAAM,MAAM,GAAG,IAAI,CAAC,UAAU,CAAC,MAAM,CAAC;QACtC,YAAY,CAAC,MAAM,CAAC,GAAG,CAAC,YAAY,CAAC,MAAM,CAAC,IAAI,CAAC,CAAC,GAAG,CAAC,CAAC;QACvD,aAAa,IAAI,IAAI,CAAC,UAAU,CAAC,UAAU,CAAC;IAC9C,CAAC;IAED,MAAM,aAAa,GAAG,KAAK,CAAC,MAAM,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,aAAa,GAAG,KAAK,CAAC,MAAM,CAAC;IAC5E,OAAO,EAAE,aAAa,EAAE,YAAY,EAAE,aAAa,EAAE,CAAC;AACxD,CAAC;AAED;;;;;GAKG;AACH,MAAM,CAAC,KAAK,UAAU,WAAW,CAAC,IAMjC;IACC,MAAM,QAAQ,GAAG,IAAI,CAAC,QAAQ,CAAC;IAC/B,MAAM,KAAK,GAAG,IAAI,CAAC,KAAK,IAAI,MAAM,CAAC;IACnC,MAAM,WAAW,GAAG,IAAI,CAAC,YAAY,CAAC;IACtC,MAAM,aAAa,GAAG,IAAI,CAAC,eAAe,IAAI,CAAC,CAAC;IAChD,MAAM,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC;IAEzB,MAAM,QAAQ,GAAG,YAAY,QAAQ,IAAI,KAAK,IAAI,WAAW,IAAI,KAAK,IAAI,aAAa,EAAE,CAAC;IAC1F,MAAM,MAAM,GAAG,KAAK,EAAE,SAAS,CAAC,QAAQ,CAAC,CAAC;IAC1C,IAAI,MAAM,KAAK,IAAI,IAAI,MAAM,KAAK,SAAS,EAAE,CAAC;QAC5C,OAAO,MAA0B,CAAC;IACpC,CAAC;IAED,kEAAkE;IAClE,MAAM,QAAQ,GAAG,aAAa,CAAC,QAAQ,CAAC,CAAC;IACzC,MAAM,OAAO,GAAG,KAAK,KAAK,OAAO,CAAC,CAAC,CAAC,QAAQ,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC,QAAQ,CAAC;IAEpE,oDAAoD;IACpD,MAAM,SAAS,GAAG,WAAW,KAAK,SAAS;QACzC,CAAC,CAAC,SAAS;QACX,CAAC,CAAC,wBAAwB,CAAC,WAAW,CAAC,CAAC;IAC1C,MAAM,OAAO,GAAG,MAAM,OAAO,CAAC,GAAG,CAC/B,OAAO,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CACpB,eAAe,CAAC,KAAK,EAAE;QACrB,KAAK,EAAE,aAAa;QACpB,SAAS;QACT,YAAY,EAAE,WAAW;QACzB,aAAa,EAAE,IAAI;KACpB,CAAC,CACH,CACF,CAAC;IAEF,yEAAyE;IACzE,MAAM,MAAM,GAAG,OAAO,CAAC,IAAI,EAAE,CAAC;IAC9B,MAAM,OAAO,GAAG,WAAW,CAAC,MAAM,CAAC,CAAC;IACpC,MAAM,MAAM,GAAG,MAAM,CAAC,OAAO,EAAE,QAAQ,EAAE,WAAW,IAAI,EAAE,CAAC,CAAC;IAE5D,iDAAiD;IACjD,MAAM,QAAQ,GAAG,WAAW,KAAK,SAAS;QACxC,CAAC,CAAC,MAAM;QACR,CAAC,CAAC,MAAM,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,YAAY,CAAC,CAAC,CAAC,SAAS,EAAE,WAAW,CAAC,CAAC,CAAC;IAEjE,MAAM,GAAG,GAAG,QAAQ,CAAC,KAAK,CAAC,CAAC,EAAE,SAAS,CAAC,CAAC;IACzC,MAAM,OAAO,GAAG,GAAG,CAAC,GAAG,CAAC,QAAQ,CAAC,CAAC;IAElC,6EAA6E;IAC7E,MAAM,KAAK,GAAG,MAAM,SAAS,CAC3B,GAAG,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,GAAG,CAAC,EACrB,KAAK,CACN,CAAC;IAEF,4EAA4E;IAC5E,wDAAwD;IACxD,MAAM,YAAY,GAAG,mBAAmB,CACtC,KAAK,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,KAAK,EAAE,EAAE,CAAC,CAAC;QAC1B,GAAG,EAAE,OAAO,CAAC,KAAK,CAAC,EAAE,GAAG,IAAI,IAAI,CAAC,GAAG;QACpC,OAAO,EAAE,IAAI,CAAC,OAAO;KACtB,CAAC,CAAC,CACJ,CAAC;IAEF,gEAAgE;IAChE,MAAM,SAAS,GAAG,gBAAgB,CAChC,KAAK,EACL,OAAO,EACP,QAAQ,EACR,YAAY,EACZ,YAAY,CACb,CAAC;IAEF,MAAM,kBAAkB,GAAG,gBAAgB,CAAC,YAAY,CAAC,CAAC;IAC1D,MAAM,QAAQ,GAAqB;QACjC,MAAM,EAAE,WAAW,CAAC,KAAK,EAAE,SAAS,CAAC;QACrC,OAAO;QACP,YAAY,EAAE,OAAO;QACrB,SAAS;QACT,QAAQ,EAAE;YACR,cAAc,EAAE,KAAK,CAAC,MAAM;YAC5B,mBAAmB,EAAE,kBAAkB;YACvC,kBAAkB,EAAE,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,KAAK,CAAC,MAAM,GAAG,kBAAkB,CAAC;YAClE,mBAAmB,EAAE,iBAAiB,CAAC,QAAQ,EAAE,SAAS,CAAC;YAC3D,WAAW,EAAE,SAAS,CAAC,MAAM;SAC9B;QACD,gBAAgB,EAAE,eAAe,CAAC,KAAK,CAAC;KACzC,CAAC;IAEF,KAAK,EAAE,SAAS,CAAC,QAAQ,EAAE,QAAQ,EAAE,QAAQ,CAAC,CAAC;IAC/C,OAAO,QAAQ,CAAC;AAClB,CAAC;AAED,sFAAsF;AACtF,MAAM,UAAU,gBAAgB,CAAC,MAAiB,EAAE,KAAa;IAC/D,MAAM,CAAC,YAAY,CACjB,cAAc,EACd;QACE,KAAK,EAAE,cAAc;QACrB,WAAW,EACT,gTAAgT;QAClT,WAAW,EAAE,mBAAmB;KACjC,EACD,KAAK,EAAE,IAAI,EAAE,EAAE;QACb,MAAM,QAAQ,GAAG,MAAM,WAAW,CAAC,EAAE,GAAG,IAAI,EAAE,KAAK,EAAE,CAAC,CAAC;QACvD,OAAO;YACL,OAAO,EAAE,CAAC,EAAE,IAAI,EAAE,MAAM,EAAE,IAAI,EAAE,IAAI,CAAC,SAAS,CAAC,QAAQ,CAAC,EAAE,CAAC;SAC5D,CAAC;IACJ,CAAC,CACF,CAAC;IAEF,MAAM,CAAC,YAAY,CACjB,aAAa,EACb;QACE,KAAK,EAAE,aAAa;QACpB,WAAW,EACT,uNAAuN;QACzN,WAAW,EAAE,mBAAmB;KACjC,EACD,KAAK,EAAE,IAAI,EAAE,EAAE;QACb,MAAM,QAAQ,GAAG,MAAM,WAAW,CAAC,EAAE,GAAG,IAAI,EAAE,KAAK,EAAE,CAAC,CAAC;QACvD,OAAO;YACL,OAAO,EAAE,CAAC,EAAE,IAAI,EAAE,MAAM,EAAE,IAAI,EAAE,IAAI,CAAC,SAAS,CAAC,QAAQ,CAAC,EAAE,CAAC;SAC5D,CAAC;IACJ,CAAC,CACF,CAAC;AACJ,CAAC"}

package/mcp.json

@@ -1,6 +1,6 @@
 {
   "name": "better-web-search-mcp",
-  "version": "0.3.0",
+  "version": "0.4.1",
   "description": "BetterWebSearch MCP — keyless-first web research (DuckDuckGo) with optional Brave/Tavily and 3-tier extraction",
   "mcpServers": {
     "better-web-search-mcp": {

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "better-web-search-mcp",
-  "version": "0.3.0",
+  "version": "0.4.1",
   "description": "MCP server for genuinely better web research — keyless-first (DuckDuckGo), 3-tier extraction, deep research. Brave/Tavily optional.",
   "type": "module",
   "bin": {
@@ -51,6 +51,7 @@
     "release:minor": "npm run version:minor && git push --follow-tags",
     "release:major": "npm run version:major && git push --follow-tags",
     "check:publish": "npm pack --dry-run",
+    "bench": "node benchmarks/run.mjs",
     "docs:api": "typedoc --options typedoc.json && node scripts/typedoc-postprocess.mjs",
     "docs:changelog": "node scripts/sync-docs-changelog.mjs",
     "docs:build": "npm run docs:api && npm run docs:changelog && hugo --source site --gc --minify",