diff --git a/examples/literature-explorer/README.md b/examples/literature-explorer/README.md
index 7ecb44a..cb19697 100644
--- a/examples/literature-explorer/README.md
+++ b/examples/literature-explorer/README.md
@@ -42,6 +42,7 @@ The adapter exposes JSON-serializable methods suitable for a local web bridge:
 - `expand_topic(...)`
 - `extract_text(text, backend="heuristic")`
 - `verify_strings(values, context="", limit=5)`
+- `support_claims(text, context="", limit=5, max_claims=8, min_claim_chars=90)`
 - `verify_bibtex(bibtex_text, context="", limit=5)`
 - `graph(seed_keys, relation_types=None, depth=1, review_status=None, missing_only=False)`
 
@@ -64,6 +65,7 @@ window.citegeist = {
   expandTopic(topicSlug, opts) {},
   extractText(text, opts) {},
   verifyStrings(values, opts) {},
+  supportClaims(text, opts) {},
   verifyBibtex(bibtexText, opts) {},
   graph(seedKeys, opts) {}
 }
@@ -122,6 +124,7 @@ This is sufficient to drive a demonstration app that can:
 - preview or apply topic expansion;
 - inspect one entry with BibTeX/provenance details;
 - run rough-reference extraction and verification;
+- rank support-worthy claims in a prose excerpt and inspect suggested references;
 - render local citation neighborhoods from `graph()` JSON payloads.
 
 For a first demo, the strongest path is topic exploration rather than generic reference-manager behavior.
diff --git a/examples/literature-explorer/index.html b/examples/literature-explorer/index.html
index bd562c3..0ce0a76 100644
--- a/examples/literature-explorer/index.html
+++ b/examples/literature-explorer/index.html
@@ -839,6 +839,23 @@
     "max_expanded_entries": 100,
     "max_expand_seconds": 20
   }
+}</div>
+          </div>
+          <div class="endpoint-card">
+            <div class="endpoint-head">
+              <span class="endpoint-method">POST</span>
+              <span class="endpoint-path">/api/call support_claims</span>
+            </div>
+            <p>Analyze a prose excerpt for support-worthy claims, rank them by <code>needs_support_score</code>, and return candidate references not already parsed from the excerpt.</p>
+            <div class="code-block">{
+  "method": "support_claims",
+  "params": {
+    "text": "Movement may not be modeled at all, but simply assigned a cost value ...",
+    "context": "artificial life",
+    "limit": 5,
+    "max_claims": 5,
+    "min_claim_chars": 80
+  }
 }</div>
           </div>
           <div class="endpoint-card">
diff --git a/tests/test_app_server.py b/tests/test_app_server.py
index 2475264..482e317 100644
--- a/tests/test_app_server.py
+++ b/tests/test_app_server.py
@@ -2,6 +2,7 @@ from types import SimpleNamespace
 
 from citegeist import BibliographyStore
 from citegeist.app_api import LiteratureExplorerApi
+from citegeist.bibtex import BibEntry
 from citegeist.bootstrap import BootstrapResult
 from citegeist.app_server import (
     LiteratureExplorerAppServer,
@@ -9,6 +10,7 @@ from citegeist.app_server import (
     _request_is_authorized,
     create_request_handler,
 )
+from citegeist.verify import VerificationMatch, VerificationResult
 
 
 class FakeBootstrapper:
@@ -29,6 +31,38 @@ class FakeBootstrapper:
         ]
 
 
+class FakeVerifier:
+    def verify_strings(self, values, context="", limit=5):
+        return []
+
+    def verify_string(self, value: str, context: str = "", limit: int = 5):
+        return VerificationResult(
+            query=value,
+            context=context,
+            status="high_confidence",
+            confidence=0.88,
+            entry=BibEntry(
+                entry_type="article",
+                citation_key="support2024",
+                fields={"title": "Support Paper", "year": "2024"},
+            ),
+            source_label="openalex:search:Support Paper",
+            alternates=[
+                VerificationMatch(
+                    entry=BibEntry(
+                        entry_type="article",
+                        citation_key="alt2023",
+                        fields={"title": "Alternate Support", "year": "2023"},
+                    ),
+                    score=0.66,
+                    source_label="crossref:search:Alternate Support",
+                )
+            ],
+            input_type="string",
+            input_key=None,
+        )
+
+
 def test_literature_explorer_app_server_dispatch_search():
     store = BibliographyStore()
     try:
@@ -138,6 +172,35 @@ def test_literature_explorer_app_server_dispatch_bootstrap_with_new_caps():
         store.close()
 
 
+def test_literature_explorer_app_server_dispatch_support_claims():
+    store = BibliographyStore()
+    try:
+        server = LiteratureExplorerAppServer(LiteratureExplorerApi(store, verifier=FakeVerifier()))
+
+        payload = server.dispatch(
+            "support_claims",
+            {
+                "text": """
+Long claim text about agents evolving intelligent movement strategies in multiple computational settings without enough direct support [1].
+
+References
+
+[[1]]Earlier Cited Paper
+""",
+                "context": "artificial life",
+                "limit": 3,
+                "max_claims": 2,
+                "min_claim_chars": 40,
+            },
+        )
+
+        assert payload["context"] == "artificial life"
+        assert payload["suggestion_count"] == 1
+        assert payload["suggestions"][0]["suggested_references"][0]["citation_key"] == "support2024"
+    finally:
+        store.close()
+
+
 def test_literature_explorer_http_handler_class_can_be_created():
     store = BibliographyStore()
     try: