diff --git a/Cargo.lock b/Cargo.lock
index 45b286d85..cdf4e699f 100644
--- a/Cargo.lock
+++ b/Cargo.lock
@@ -3584,8 +3584,8 @@ checksum = "42703706b716c37f96a77aea830392ad231f44c9e9a67872fa5548707e11b11c"
 
 [[package]]
 name = "fsst"
-version = "14.0.0-beta.11"
-source = "git+https://github.com/lance-format/lance.git?tag=v14.0.0-beta.11#f9c563fe1516a48b5c40d255c1d884af2f3c3d6c"
+version = "14.0.0-beta.12"
+source = "git+https://github.com/lance-format/lance.git?tag=v14.0.0-beta.12#b3c8619886b9ba3a26f323241d16c7f36b28db7f"
 dependencies = [
  "arrow-array",
  "rand 0.9.5",
@@ -5059,8 +5059,8 @@ checksum = "e037a2e1d8d5fdbd49b16a4ea09d5d6401c1f29eca5ff29d03d3824dba16256a"
 
 [[package]]
 name = "lance"
-version = "14.0.0-beta.11"
-source = "git+https://github.com/lance-format/lance.git?tag=v14.0.0-beta.11#f9c563fe1516a48b5c40d255c1d884af2f3c3d6c"
+version = "14.0.0-beta.12"
+source = "git+https://github.com/lance-format/lance.git?tag=v14.0.0-beta.12#b3c8619886b9ba3a26f323241d16c7f36b28db7f"
 dependencies = [
  "arc-swap",
  "arrow",
@@ -5133,8 +5133,8 @@ dependencies = [
 
 [[package]]
 name = "lance-arrow"
-version = "14.0.0-beta.11"
-source = "git+https://github.com/lance-format/lance.git?tag=v14.0.0-beta.11#f9c563fe1516a48b5c40d255c1d884af2f3c3d6c"
+version = "14.0.0-beta.12"
+source = "git+https://github.com/lance-format/lance.git?tag=v14.0.0-beta.12#b3c8619886b9ba3a26f323241d16c7f36b28db7f"
 dependencies = [
  "arrow-array",
  "arrow-buffer",
@@ -5156,7 +5156,7 @@ dependencies = [
 [[package]]
 name = "lance-arrow-scalar"
 version = "58.0.0"
-source = "git+https://github.com/lance-format/lance.git?tag=v14.0.0-beta.11#f9c563fe1516a48b5c40d255c1d884af2f3c3d6c"
+source = "git+https://github.com/lance-format/lance.git?tag=v14.0.0-beta.12#b3c8619886b9ba3a26f323241d16c7f36b28db7f"
 dependencies = [
  "arrow-array",
  "arrow-buffer",
@@ -5170,7 +5170,7 @@ dependencies = [
 [[package]]
 name = "lance-arrow-stats"
 version = "58.0.0"
-source = "git+https://github.com/lance-format/lance.git?tag=v14.0.0-beta.11#f9c563fe1516a48b5c40d255c1d884af2f3c3d6c"
+source = "git+https://github.com/lance-format/lance.git?tag=v14.0.0-beta.12#b3c8619886b9ba3a26f323241d16c7f36b28db7f"
 dependencies = [
  "arrow-array",
  "arrow-schema",
@@ -5179,8 +5179,8 @@ dependencies = [
 
 [[package]]
 name = "lance-bitpacking"
-version = "14.0.0-beta.11"
-source = "git+https://github.com/lance-format/lance.git?tag=v14.0.0-beta.11#f9c563fe1516a48b5c40d255c1d884af2f3c3d6c"
+version = "14.0.0-beta.12"
+source = "git+https://github.com/lance-format/lance.git?tag=v14.0.0-beta.12#b3c8619886b9ba3a26f323241d16c7f36b28db7f"
 dependencies = [
  "arrayref",
  "crunchy",
@@ -5190,8 +5190,8 @@ dependencies = [
 
 [[package]]
 name = "lance-core"
-version = "14.0.0-beta.11"
-source = "git+https://github.com/lance-format/lance.git?tag=v14.0.0-beta.11#f9c563fe1516a48b5c40d255c1d884af2f3c3d6c"
+version = "14.0.0-beta.12"
+source = "git+https://github.com/lance-format/lance.git?tag=v14.0.0-beta.12#b3c8619886b9ba3a26f323241d16c7f36b28db7f"
 dependencies = [
  "arrow-array",
  "arrow-buffer",
@@ -5228,8 +5228,8 @@ dependencies = [
 
 [[package]]
 name = "lance-datafusion"
-version = "14.0.0-beta.11"
-source = "git+https://github.com/lance-format/lance.git?tag=v14.0.0-beta.11#f9c563fe1516a48b5c40d255c1d884af2f3c3d6c"
+version = "14.0.0-beta.12"
+source = "git+https://github.com/lance-format/lance.git?tag=v14.0.0-beta.12#b3c8619886b9ba3a26f323241d16c7f36b28db7f"
 dependencies = [
  "arrow",
  "arrow-array",
@@ -5260,8 +5260,8 @@ dependencies = [
 
 [[package]]
 name = "lance-datagen"
-version = "14.0.0-beta.11"
-source = "git+https://github.com/lance-format/lance.git?tag=v14.0.0-beta.11#f9c563fe1516a48b5c40d255c1d884af2f3c3d6c"
+version = "14.0.0-beta.12"
+source = "git+https://github.com/lance-format/lance.git?tag=v14.0.0-beta.12#b3c8619886b9ba3a26f323241d16c7f36b28db7f"
 dependencies = [
  "arrow",
  "arrow-array",
@@ -5278,8 +5278,8 @@ dependencies = [
 
 [[package]]
 name = "lance-derive"
-version = "14.0.0-beta.11"
-source = "git+https://github.com/lance-format/lance.git?tag=v14.0.0-beta.11#f9c563fe1516a48b5c40d255c1d884af2f3c3d6c"
+version = "14.0.0-beta.12"
+source = "git+https://github.com/lance-format/lance.git?tag=v14.0.0-beta.12#b3c8619886b9ba3a26f323241d16c7f36b28db7f"
 dependencies = [
  "proc-macro2",
  "quote",
@@ -5288,8 +5288,8 @@ dependencies = [
 
 [[package]]
 name = "lance-encoding"
-version = "14.0.0-beta.11"
-source = "git+https://github.com/lance-format/lance.git?tag=v14.0.0-beta.11#f9c563fe1516a48b5c40d255c1d884af2f3c3d6c"
+version = "14.0.0-beta.12"
+source = "git+https://github.com/lance-format/lance.git?tag=v14.0.0-beta.12#b3c8619886b9ba3a26f323241d16c7f36b28db7f"
 dependencies = [
  "arrow-arith",
  "arrow-array",
@@ -5322,8 +5322,8 @@ dependencies = [
 
 [[package]]
 name = "lance-file"
-version = "14.0.0-beta.11"
-source = "git+https://github.com/lance-format/lance.git?tag=v14.0.0-beta.11#f9c563fe1516a48b5c40d255c1d884af2f3c3d6c"
+version = "14.0.0-beta.12"
+source = "git+https://github.com/lance-format/lance.git?tag=v14.0.0-beta.12#b3c8619886b9ba3a26f323241d16c7f36b28db7f"
 dependencies = [
  "arrow-arith",
  "arrow-array",
@@ -5355,8 +5355,8 @@ dependencies = [
 
 [[package]]
 name = "lance-geo"
-version = "14.0.0-beta.11"
-source = "git+https://github.com/lance-format/lance.git?tag=v14.0.0-beta.11#f9c563fe1516a48b5c40d255c1d884af2f3c3d6c"
+version = "14.0.0-beta.12"
+source = "git+https://github.com/lance-format/lance.git?tag=v14.0.0-beta.12#b3c8619886b9ba3a26f323241d16c7f36b28db7f"
 dependencies = [
  "datafusion",
  "geo-traits",
@@ -5370,8 +5370,8 @@ dependencies = [
 
 [[package]]
 name = "lance-index"
-version = "14.0.0-beta.11"
-source = "git+https://github.com/lance-format/lance.git?tag=v14.0.0-beta.11#f9c563fe1516a48b5c40d255c1d884af2f3c3d6c"
+version = "14.0.0-beta.12"
+source = "git+https://github.com/lance-format/lance.git?tag=v14.0.0-beta.12#b3c8619886b9ba3a26f323241d16c7f36b28db7f"
 dependencies = [
  "arc-swap",
  "arrow",
@@ -5442,8 +5442,8 @@ dependencies = [
 
 [[package]]
 name = "lance-index-core"
-version = "14.0.0-beta.11"
-source = "git+https://github.com/lance-format/lance.git?tag=v14.0.0-beta.11#f9c563fe1516a48b5c40d255c1d884af2f3c3d6c"
+version = "14.0.0-beta.12"
+source = "git+https://github.com/lance-format/lance.git?tag=v14.0.0-beta.12#b3c8619886b9ba3a26f323241d16c7f36b28db7f"
 dependencies = [
  "arrow-array",
  "arrow-schema",
@@ -5466,8 +5466,8 @@ dependencies = [
 
 [[package]]
 name = "lance-io"
-version = "14.0.0-beta.11"
-source = "git+https://github.com/lance-format/lance.git?tag=v14.0.0-beta.11#f9c563fe1516a48b5c40d255c1d884af2f3c3d6c"
+version = "14.0.0-beta.12"
+source = "git+https://github.com/lance-format/lance.git?tag=v14.0.0-beta.12#b3c8619886b9ba3a26f323241d16c7f36b28db7f"
 dependencies = [
  "arrow",
  "arrow-array",
@@ -5509,8 +5509,8 @@ dependencies = [
 
 [[package]]
 name = "lance-jsonb"
-version = "14.0.0-beta.11"
-source = "git+https://github.com/lance-format/lance.git?tag=v14.0.0-beta.11#f9c563fe1516a48b5c40d255c1d884af2f3c3d6c"
+version = "14.0.0-beta.12"
+source = "git+https://github.com/lance-format/lance.git?tag=v14.0.0-beta.12#b3c8619886b9ba3a26f323241d16c7f36b28db7f"
 dependencies = [
  "byteorder",
  "ethnum",
@@ -5527,8 +5527,8 @@ dependencies = [
 
 [[package]]
 name = "lance-linalg"
-version = "14.0.0-beta.11"
-source = "git+https://github.com/lance-format/lance.git?tag=v14.0.0-beta.11#f9c563fe1516a48b5c40d255c1d884af2f3c3d6c"
+version = "14.0.0-beta.12"
+source = "git+https://github.com/lance-format/lance.git?tag=v14.0.0-beta.12#b3c8619886b9ba3a26f323241d16c7f36b28db7f"
 dependencies = [
  "arrow-array",
  "arrow-schema",
@@ -5542,8 +5542,8 @@ dependencies = [
 
 [[package]]
 name = "lance-namespace"
-version = "14.0.0-beta.11"
-source = "git+https://github.com/lance-format/lance.git?tag=v14.0.0-beta.11#f9c563fe1516a48b5c40d255c1d884af2f3c3d6c"
+version = "14.0.0-beta.12"
+source = "git+https://github.com/lance-format/lance.git?tag=v14.0.0-beta.12#b3c8619886b9ba3a26f323241d16c7f36b28db7f"
 dependencies = [
  "arrow",
  "async-trait",
@@ -5557,8 +5557,8 @@ dependencies = [
 
 [[package]]
 name = "lance-namespace-impls"
-version = "14.0.0-beta.11"
-source = "git+https://github.com/lance-format/lance.git?tag=v14.0.0-beta.11#f9c563fe1516a48b5c40d255c1d884af2f3c3d6c"
+version = "14.0.0-beta.12"
+source = "git+https://github.com/lance-format/lance.git?tag=v14.0.0-beta.12#b3c8619886b9ba3a26f323241d16c7f36b28db7f"
 dependencies = [
  "arc-swap",
  "arrow",
@@ -5612,8 +5612,8 @@ dependencies = [
 
 [[package]]
 name = "lance-select"
-version = "14.0.0-beta.11"
-source = "git+https://github.com/lance-format/lance.git?tag=v14.0.0-beta.11#f9c563fe1516a48b5c40d255c1d884af2f3c3d6c"
+version = "14.0.0-beta.12"
+source = "git+https://github.com/lance-format/lance.git?tag=v14.0.0-beta.12#b3c8619886b9ba3a26f323241d16c7f36b28db7f"
 dependencies = [
  "arrow-array",
  "arrow-buffer",
@@ -5627,8 +5627,8 @@ dependencies = [
 
 [[package]]
 name = "lance-table"
-version = "14.0.0-beta.11"
-source = "git+https://github.com/lance-format/lance.git?tag=v14.0.0-beta.11#f9c563fe1516a48b5c40d255c1d884af2f3c3d6c"
+version = "14.0.0-beta.12"
+source = "git+https://github.com/lance-format/lance.git?tag=v14.0.0-beta.12#b3c8619886b9ba3a26f323241d16c7f36b28db7f"
 dependencies = [
  "arrow",
  "arrow-array",
@@ -5668,8 +5668,8 @@ dependencies = [
 
 [[package]]
 name = "lance-testing"
-version = "14.0.0-beta.11"
-source = "git+https://github.com/lance-format/lance.git?tag=v14.0.0-beta.11#f9c563fe1516a48b5c40d255c1d884af2f3c3d6c"
+version = "14.0.0-beta.12"
+source = "git+https://github.com/lance-format/lance.git?tag=v14.0.0-beta.12#b3c8619886b9ba3a26f323241d16c7f36b28db7f"
 dependencies = [
  "arrow-array",
  "arrow-schema",
@@ -5682,8 +5682,8 @@ dependencies = [
 
 [[package]]
 name = "lance-tokenizer"
-version = "14.0.0-beta.11"
-source = "git+https://github.com/lance-format/lance.git?tag=v14.0.0-beta.11#f9c563fe1516a48b5c40d255c1d884af2f3c3d6c"
+version = "14.0.0-beta.12"
+source = "git+https://github.com/lance-format/lance.git?tag=v14.0.0-beta.12#b3c8619886b9ba3a26f323241d16c7f36b28db7f"
 dependencies = [
  "frostem",
  "icu_segmenter",
diff --git a/Cargo.toml b/Cargo.toml
index 49d1792de..9246f85fc 100644
--- a/Cargo.toml
+++ b/Cargo.toml
@@ -13,20 +13,20 @@ categories = ["database-implementations"]
 rust-version = "1.94.0"
 
 [workspace.dependencies]
-lance = { "version" = "=14.0.0-beta.11", default-features = false, "tag" = "v14.0.0-beta.11", "git" = "https://github.com/lance-format/lance.git" }
-lance-core = { "version" = "=14.0.0-beta.11", "tag" = "v14.0.0-beta.11", "git" = "https://github.com/lance-format/lance.git" }
-lance-datagen = { "version" = "=14.0.0-beta.11", "tag" = "v14.0.0-beta.11", "git" = "https://github.com/lance-format/lance.git" }
-lance-file = { "version" = "=14.0.0-beta.11", "tag" = "v14.0.0-beta.11", "git" = "https://github.com/lance-format/lance.git" }
-lance-io = { "version" = "=14.0.0-beta.11", default-features = false, "tag" = "v14.0.0-beta.11", "git" = "https://github.com/lance-format/lance.git" }
-lance-index = { "version" = "=14.0.0-beta.11", "tag" = "v14.0.0-beta.11", "git" = "https://github.com/lance-format/lance.git" }
-lance-linalg = { "version" = "=14.0.0-beta.11", "tag" = "v14.0.0-beta.11", "git" = "https://github.com/lance-format/lance.git" }
-lance-namespace = { "version" = "=14.0.0-beta.11", "tag" = "v14.0.0-beta.11", "git" = "https://github.com/lance-format/lance.git" }
-lance-namespace-impls = { "version" = "=14.0.0-beta.11", default-features = false, "tag" = "v14.0.0-beta.11", "git" = "https://github.com/lance-format/lance.git" }
-lance-table = { "version" = "=14.0.0-beta.11", "tag" = "v14.0.0-beta.11", "git" = "https://github.com/lance-format/lance.git" }
-lance-testing = { "version" = "=14.0.0-beta.11", "tag" = "v14.0.0-beta.11", "git" = "https://github.com/lance-format/lance.git" }
-lance-datafusion = { "version" = "=14.0.0-beta.11", "tag" = "v14.0.0-beta.11", "git" = "https://github.com/lance-format/lance.git" }
-lance-encoding = { "version" = "=14.0.0-beta.11", "tag" = "v14.0.0-beta.11", "git" = "https://github.com/lance-format/lance.git" }
-lance-arrow = { "version" = "=14.0.0-beta.11", "tag" = "v14.0.0-beta.11", "git" = "https://github.com/lance-format/lance.git" }
+lance = { "version" = "=14.0.0-beta.12", default-features = false, "tag" = "v14.0.0-beta.12", "git" = "https://github.com/lance-format/lance.git" }
+lance-core = { "version" = "=14.0.0-beta.12", "tag" = "v14.0.0-beta.12", "git" = "https://github.com/lance-format/lance.git" }
+lance-datagen = { "version" = "=14.0.0-beta.12", "tag" = "v14.0.0-beta.12", "git" = "https://github.com/lance-format/lance.git" }
+lance-file = { "version" = "=14.0.0-beta.12", "tag" = "v14.0.0-beta.12", "git" = "https://github.com/lance-format/lance.git" }
+lance-io = { "version" = "=14.0.0-beta.12", default-features = false, "tag" = "v14.0.0-beta.12", "git" = "https://github.com/lance-format/lance.git" }
+lance-index = { "version" = "=14.0.0-beta.12", "tag" = "v14.0.0-beta.12", "git" = "https://github.com/lance-format/lance.git" }
+lance-linalg = { "version" = "=14.0.0-beta.12", "tag" = "v14.0.0-beta.12", "git" = "https://github.com/lance-format/lance.git" }
+lance-namespace = { "version" = "=14.0.0-beta.12", "tag" = "v14.0.0-beta.12", "git" = "https://github.com/lance-format/lance.git" }
+lance-namespace-impls = { "version" = "=14.0.0-beta.12", default-features = false, "tag" = "v14.0.0-beta.12", "git" = "https://github.com/lance-format/lance.git" }
+lance-table = { "version" = "=14.0.0-beta.12", "tag" = "v14.0.0-beta.12", "git" = "https://github.com/lance-format/lance.git" }
+lance-testing = { "version" = "=14.0.0-beta.12", "tag" = "v14.0.0-beta.12", "git" = "https://github.com/lance-format/lance.git" }
+lance-datafusion = { "version" = "=14.0.0-beta.12", "tag" = "v14.0.0-beta.12", "git" = "https://github.com/lance-format/lance.git" }
+lance-encoding = { "version" = "=14.0.0-beta.12", "tag" = "v14.0.0-beta.12", "git" = "https://github.com/lance-format/lance.git" }
+lance-arrow = { "version" = "=14.0.0-beta.12", "tag" = "v14.0.0-beta.12", "git" = "https://github.com/lance-format/lance.git" }
 lancedb = { path = "rust/lancedb", default-features = false }
 ahash = "0.8"
 # Note that this one does not include pyarrow
diff --git a/nodejs/__test__/query.test.ts b/nodejs/__test__/query.test.ts
index 16a595002..786f04015 100644
--- a/nodejs/__test__/query.test.ts
+++ b/nodejs/__test__/query.test.ts
@@ -527,3 +527,106 @@ describe("Query orderBy", () => {
     ]);
   });
 });
+
+describe("IVF search effort", () => {
+  let directory: tmp.DirResult;
+  let table: Table;
+  const vector = Array.from({ length: 16 }, (_, i) => (i === 0 ? 1 : 0));
+
+  beforeEach(async () => {
+    directory = tmp.dirSync({ unsafeCleanup: true });
+    const db = await connect(directory.name);
+    const rows = Array.from({ length: 512 }, (_, id) => ({
+      id,
+      vector: Array.from(
+        { length: 16 },
+        (_, j) =>
+          (Math.floor(id / 32) === j ? 1 : 0) + Math.sin(id * 17 + j) * 0.001,
+      ),
+    }));
+    table = await db.createTable("effort", rows.slice(0, 256));
+    await table.add(rows.slice(256));
+    await table.createIndex("vector", {
+      config: Index.ivfFlat({ numPartitions: 16 }),
+    });
+  });
+
+  afterEach(() => {
+    directory.removeCallback();
+  });
+
+  it.each([
+    [null, 6],
+    [0, 1],
+    [0.25, 3],
+    [0.5, 6],
+    [0.75, 10],
+    [1, 16],
+  ])(
+    "preserves effort %s through execution and probe precedence",
+    async (effort, expected) => {
+      const partitions = async (query: ReturnType<Table["vectorSearch"]>) => {
+        const plan = await query.limit(8).analyzePlan();
+        const matches = [...plan.matchAll(/partitions_searched=(\d+)/g)];
+        expect(matches.length).toBeGreaterThan(0);
+        return matches.reduce((total, match) => total + Number(match[1]), 0);
+      };
+      expect(
+        await partitions(table.vectorSearch(vector).searchEffort(effort)),
+      ).toBe(expected);
+      expect(
+        await partitions(
+          table.vectorSearch(vector).nprobes(1).searchEffort(effort),
+        ),
+      ).toBe(effort === null ? 1 : expected);
+      expect(
+        await partitions(
+          table.vectorSearch(vector).searchEffort(effort).nprobes(1),
+        ),
+      ).toBe(effort === null ? 1 : expected);
+      expect(
+        await partitions(
+          table
+            .vectorSearch(vector)
+            .searchEffort(effort)
+            .nprobes(1)
+            .searchEffort(null),
+        ),
+      ).toBe(1);
+      expect(
+        await partitions(
+          table.vectorSearch(vector).searchEffort(effort).maximumNprobes(2),
+        ),
+      ).toBeLessThanOrEqual(2);
+      const actual = await table
+        .vectorSearch(vector)
+        .searchEffort(effort)
+        .limit(8)
+        .toArray();
+      const exact = await table
+        .vectorSearch(vector)
+        .bypassVectorIndex()
+        .limit(8)
+        .toArray();
+      const ids = new Set(exact.map((row) => row.id));
+      expect(
+        actual.filter((row) => ids.has(row.id)).length / 8,
+      ).toBeGreaterThanOrEqual(0.5);
+    },
+  );
+
+  it.each([
+    -0.1,
+    1.1,
+    Number.NaN,
+    Number.POSITIVE_INFINITY,
+    Number.NEGATIVE_INFINITY,
+  ])("rejects invalid effort %s before execution", (effort) => {
+    expect(() => table.vectorSearch(vector).searchEffort(effort)).toThrow(
+      "search_effort must be finite and in [0, 1]",
+    );
+    expect(() =>
+      table.vectorSearch(vector).fullTextSearch("dog").searchEffort(effort),
+    ).toThrow("search_effort must be finite and in [0, 1]");
+  });
+});
diff --git a/nodejs/lancedb/query.ts b/nodejs/lancedb/query.ts
index 0f0eaf610..cce505c97 100644
--- a/nodejs/lancedb/query.ts
+++ b/nodejs/lancedb/query.ts
@@ -549,6 +549,27 @@ export class VectorQuery extends StandardQueryBase<NativeVectorQuery> {
     return this;
   }
 
+  /**
+   * Set the initial IVF search effort to a finite value in [0, 1].
+   *
+   * The default, null, uses Auto (internally 0.5). Zero starts at the caller
+   * minimum; one searches all partitions within the caller maximum. Later
+   * probing may continue when too few candidates are found.
+   * Any explicit effort, including 0.5, overrides nprobes in either call order.
+   * Explicit minimum and maximum bounds still apply. This is not a recall
+   * guarantee and does not change refinement or HNSW ef. The MemWAL LSM
+   * scanner does not yet support this option.
+   *
+   * @example
+   * ```ts
+   * table.vectorSearch([0.1, 0.2]).searchEffort(0.75).limit(10);
+   * ```
+   */
+  searchEffort(searchEffort: number | null): VectorQuery {
+    this.doVectorCall((inner) => inner.searchEffort(searchEffort));
+    return this;
+  }
+
   /**
    * Set the minimum number of probes used.
    *
diff --git a/nodejs/src/query.rs b/nodejs/src/query.rs
index 712569b2f..480426411 100644
--- a/nodejs/src/query.rs
+++ b/nodejs/src/query.rs
@@ -298,6 +298,16 @@ impl VectorQuery {
         Ok(())
     }
 
+    #[napi]
+    pub fn search_effort(&mut self, search_effort: Option<f64>) -> napi::Result<()> {
+        self.inner = self
+            .inner
+            .clone()
+            .search_effort(search_effort)
+            .default_error()?;
+        Ok(())
+    }
+
     #[napi]
     pub fn minimum_nprobes(&mut self, minimum_nprobe: u32) -> napi::Result<()> {
         self.inner = self
diff --git a/python/pyproject.toml b/python/pyproject.toml
index cde7247f3..ed54ae536 100644
--- a/python/pyproject.toml
+++ b/python/pyproject.toml
@@ -63,7 +63,7 @@ tests = [
     "polars>=0.19, <=1.32.3",
     "pyarrow<25",
     "pyarrow-stubs>=16.0",
-    "pylance==14.0.0b10",
+    "pylance==14.0.0b12",
     "requests>=2.31.0",
     "datafusion>=54,<55",
     "opentelemetry-sdk>=1.30.0",
@@ -73,6 +73,7 @@ dev = [
     "pre-commit>=3.5.0",
     "pyright>=1.1.350",
     'typing-extensions>=4.0.0; python_version < "3.11"',
+    "maturin>=1.10",
 ]
 docs = ["mkdocs", "mkdocs-jupyter", "mkdocs-material", "mkdocstrings-python"]
 clip = ["torch", "pillow>=12.1.1", "open-clip-torch"]
diff --git a/python/python/lancedb/_lancedb.pyi b/python/python/lancedb/_lancedb.pyi
index cfaa8c68a..8d0a86e4a 100644
--- a/python/python/lancedb/_lancedb.pyi
+++ b/python/python/lancedb/_lancedb.pyi
@@ -727,6 +727,7 @@ class VectorQuery:
     def use_lsm(self, enable: bool): ...
     def refine_factor(self, refine_factor: int): ...
     def nprobes(self, nprobes: int): ...
+    def search_effort(self, search_effort: Optional[float]): ...
     def minimum_nprobes(self, minimum_nprobes: int): ...
     def maximum_nprobes(self, maximum_nprobes: int): ...
     def bypass_vector_index(self): ...
@@ -752,6 +753,7 @@ class HybridQuery:
     def distance_type(self, distance_type: str): ...
     def refine_factor(self, refine_factor: int): ...
     def nprobes(self, nprobes: int): ...
+    def search_effort(self, search_effort: Optional[float]): ...
     def minimum_nprobes(self, minimum_nprobes: int): ...
     def maximum_nprobes(self, maximum_nprobes: int): ...
     def bypass_vector_index(self): ...
@@ -781,6 +783,7 @@ class PyQueryRequest:
     nprobes: Optional[int]
     minimum_nprobes: Optional[int]
     maximum_nprobes: Optional[int]
+    search_effort: Optional[float]
     lower_bound: Optional[float]
     upper_bound: Optional[float]
     ef: Optional[int]
diff --git a/python/python/lancedb/namespace.py b/python/python/lancedb/namespace.py
index ed87e441f..26c62925f 100644
--- a/python/python/lancedb/namespace.py
+++ b/python/python/lancedb/namespace.py
@@ -105,6 +105,11 @@ def _query_to_namespace_request(
 
     if not isinstance(query, Query):
         raise TypeError(f"Expected Query, got {type(query)}")
+    if query.search_effort is not None:
+        raise ValueError(
+            "namespace QueryTable does not support search_effort; "
+            "execute the query locally"
+        )
 
     vector_request = None
     if query.vector is not None:
diff --git a/python/python/lancedb/query.py b/python/python/lancedb/query.py
index c7664ab78..fc9aea9a4 100644
--- a/python/python/lancedb/query.py
+++ b/python/python/lancedb/query.py
@@ -475,8 +475,8 @@ class MatchQuery(FullTextQuery):
     ----------
     query : str
         The query string to match against.
-    column : str
-        The name of the column to match against.
+    column : str, optional
+        The column to match against. None leaves column selection to query planning.
     boost : float, default 1.0
         The boost factor for the query.
         The score of each matching document is multiplied by this value.
@@ -505,7 +505,7 @@ class MatchQuery(FullTextQuery):
     """
 
     query: str
-    column: str
+    column: Optional[str] = None
     boost: float = pydantic.Field(1.0, kw_only=True)
     fuzziness: Optional[int] = pydantic.Field(0, kw_only=True)
     max_expansions: int = pydantic.Field(50, kw_only=True)
@@ -793,6 +793,9 @@ class Query(pydantic.BaseModel):
     # and all partitions could be searched if needed to satisfy the limit.
     maximum_nprobes: Optional[int] = None
 
+    # Optional initial IVF effort; None preserves Auto and explicit nprobes.
+    search_effort: Optional[float] = None
+
     # lower bound for distance search
     lower_bound: Optional[float] = None
 
@@ -854,6 +857,7 @@ class Query(pydantic.BaseModel):
         query.nprobes = req.nprobes
         query.minimum_nprobes = req.minimum_nprobes
         query.maximum_nprobes = req.maximum_nprobes
+        query.search_effort = req.search_effort
         query.lower_bound = req.lower_bound
         query.upper_bound = req.upper_bound
         query.ef = req.ef
@@ -1635,6 +1639,7 @@ class LanceVectorQueryBuilder(LanceQueryBuilder):
         self._nprobes = None
         self._minimum_nprobes = None
         self._maximum_nprobes = None
+        self._search_effort = None
         self._lower_bound = None
         self._upper_bound = None
         self._refine_factor = None
@@ -1716,6 +1721,24 @@ class LanceVectorQueryBuilder(LanceQueryBuilder):
         self._nprobes = nprobes
         return self
 
+    def search_effort(self, search_effort: Optional[float]) -> LanceVectorQueryBuilder:
+        """Set the initial IVF search effort to a finite value in [0, 1].
+
+        None (the default) uses Auto, internally 0.5. Zero starts at the caller
+        minimum; one searches all partitions within the caller maximum. Later
+        probing may continue when too few candidates are found. Any explicit
+        effort, including 0.5, overrides nprobes in either call order. Explicit
+        minimum and maximum bounds still apply. This is not a recall guarantee
+        and does not change refinement or HNSW ef. MemWAL LSM queries do not
+        yet support this option.
+
+        Examples
+        --------
+        >>> query = table.search([0.1, 0.2]).search_effort(0.75)
+        """
+        self._search_effort = search_effort
+        return self
+
     def minimum_nprobes(self, minimum_nprobes: int) -> LanceVectorQueryBuilder:
         """Set the minimum number of probes to use.
 
@@ -1858,6 +1881,7 @@ class LanceVectorQueryBuilder(LanceQueryBuilder):
             nprobes=self._nprobes,
             minimum_nprobes=self._minimum_nprobes,
             maximum_nprobes=self._maximum_nprobes,
+            search_effort=self._search_effort,
             lower_bound=self._lower_bound,
             upper_bound=self._upper_bound,
             refine_factor=self._refine_factor,
@@ -2217,6 +2241,7 @@ class LanceHybridQueryBuilder(LanceQueryBuilder):
         self._nprobes = None
         self._minimum_nprobes = None
         self._maximum_nprobes = None
+        self._search_effort = None
         self._refine_factor = None
         self._distance_type = None
         self._phrase_query = None
@@ -2524,6 +2549,26 @@ class LanceHybridQueryBuilder(LanceQueryBuilder):
         self._nprobes = nprobes
         return self
 
+    def search_effort(self, search_effort: Optional[float]) -> LanceHybridQueryBuilder:
+        """Set the initial IVF search effort to a finite value in [0, 1].
+
+        None (the default) uses Auto, internally 0.5. Zero starts at the caller
+        minimum; one searches all partitions within the caller maximum. Later
+        probing may continue when too few candidates are found. Any explicit
+        effort, including 0.5, overrides nprobes in either call order. Explicit
+        minimum and maximum bounds still apply. This is not a recall guarantee
+        and does not change refinement or HNSW ef. MemWAL LSM queries do not
+        yet support this option.
+
+        Examples
+        --------
+        >>> query = table.search([0.1, 0.2]).search_effort(0.75)
+        """
+        self._search_effort = search_effort
+        if self._vector_query is not None:
+            self._vector_query.search_effort(search_effort)
+        return self
+
     def minimum_nprobes(self, minimum_nprobes: int) -> LanceHybridQueryBuilder:
         """Set the minimum number of probes to use.
 
@@ -2802,6 +2847,8 @@ class LanceHybridQueryBuilder(LanceQueryBuilder):
             self._vector_query.metric(self._distance_type)
         if self._nprobes is not None:
             self._vector_query.nprobes(self._nprobes)
+        if self._search_effort is not None:
+            self._vector_query.search_effort(self._search_effort)
         if self._minimum_nprobes is not None:
             self._vector_query.minimum_nprobes(self._minimum_nprobes)
         if self._maximum_nprobes is not None:
@@ -3730,6 +3777,24 @@ class AsyncVectorQueryBase:
         self._inner.nprobes(nprobes)
         return self
 
+    def search_effort(self, search_effort: Optional[float]) -> Self:
+        """Set the initial IVF search effort to a finite value in [0, 1].
+
+        None preserves Auto (internally 0.5) and fixed nprobes. Any explicit
+        effort, including 0.5, overrides nprobes in either call order; explicit
+        minimum and maximum bounds remain effective. Zero starts at the caller
+        minimum and one searches all partitions within the maximum. Later
+        probing may continue to obtain enough candidates. This is not a recall
+        guarantee and does not change refinement or HNSW ef. MemWAL LSM queries
+        do not yet support this option.
+
+        Examples
+        --------
+        >>> query = table.vector_search([0.1, 0.2]).search_effort(0.75)
+        """
+        self._inner.search_effort(search_effort)
+        return self
+
     def minimum_nprobes(self, minimum_nprobes: int) -> Self:
         """Set the minimum number of probes to use.
 
diff --git a/python/python/lancedb/table.py b/python/python/lancedb/table.py
index 21838ecef..81b376ebd 100644
--- a/python/python/lancedb/table.py
+++ b/python/python/lancedb/table.py
@@ -4440,6 +4440,7 @@ class LanceTable(Table):
             and not self._route_pushdown_to_rust
             and self.current_branch() is None
             and query.take_offsets is None
+            and query.search_effort is None
         ):
             from lancedb.namespace import _execute_server_side_query
 
@@ -6441,6 +6442,8 @@ class AsyncTable:
                 async_query = async_query.distance_type(query.distance_type)
             if query.nprobes is not None:
                 async_query = async_query.nprobes(query.nprobes)
+            if query.search_effort is not None:
+                async_query = async_query.search_effort(query.search_effort)
             if query.minimum_nprobes is not None and query.maximum_nprobes is not None:
                 async_query = async_query.minimum_nprobes(
                     query.minimum_nprobes
@@ -6481,6 +6484,7 @@ class AsyncTable:
             )
             and not self._route_pushdown_to_rust
             and query.take_offsets is None
+            and query.search_effort is None
         ):
             from lancedb.namespace import _execute_server_side_query
 
diff --git a/python/python/tests/test_namespace.py b/python/python/tests/test_namespace.py
index 76aa5fc55..520a6992d 100644
--- a/python/python/tests/test_namespace.py
+++ b/python/python/tests/test_namespace.py
@@ -1209,3 +1209,15 @@ def test_local_table_to_arrow_and_to_pandas_are_unchanged(tmp_path):
 
     assert table.to_arrow().column("id").to_pylist() == [1, 2]
     assert table.to_pandas()["id"].tolist() == [1, 2]
+
+
+@pytest.mark.parametrize("effort", [0.0, 0.5, 1.0])
+def test_namespace_request_rejects_search_effort(effort):
+    from lancedb.query import Query
+
+    with pytest.raises(
+        ValueError, match="namespace QueryTable does not support search_effort"
+    ):
+        _query_to_namespace_request(
+            ["test"], Query(vector=[1.0, 0.0], search_effort=effort)
+        )
diff --git a/python/python/tests/test_query.py b/python/python/tests/test_query.py
index daee0dd07..71c293628 100644
--- a/python/python/tests/test_query.py
+++ b/python/python/tests/test_query.py
@@ -2265,3 +2265,112 @@ def test_blob_v2_to_batches_row_id(tmp_db):
     assert "_rowid" in hits.column_names
     blobs = table.fetch_blobs("blob", hits)
     assert [blobs[i].as_py() for i in range(len(blobs))] == [b"one", b"two"]
+
+
+@pytest.mark.parametrize("metric", ["l2", "cosine", "dot"])
+@pytest.mark.parametrize(
+    "effort,probes", [(None, 4), (0, 1), (0.25, 2), (0.5, 4), (0.75, 8), (1, 16)]
+)
+def test_search_effort_probe_budget(tmp_path, monkeypatch, metric, effort, probes):
+    import re
+    from lancedb.index import IvfFlat
+
+    monkeypatch.setenv("LANCE_AUTO_PROBE_MARGIN", "0")
+    monkeypatch.setenv("LANCE_AUTO_MIN_INITIAL_NPROBES", "4")
+    monkeypatch.setenv("LANCE_AUTO_MAX_INITIAL_NPROBES", "4")
+    rng = np.random.default_rng(42)
+    vectors = np.repeat(np.eye(16, dtype=np.float32), 32, axis=0)
+    vectors += rng.normal(0, 0.001, vectors.shape).astype(np.float32)
+    data = pa.table(
+        {
+            "id": np.arange(512),
+            "vector": pa.FixedSizeListArray.from_arrays(pa.array(vectors.ravel()), 16),
+        }
+    )
+    table = lancedb.connect(tmp_path).create_table("effort", data.slice(0, 256))
+    table.add(data.slice(256))
+    table.create_index(
+        "vector", config=IvfFlat(distance_type=metric, num_partitions=16)
+    )
+    vector = np.eye(16, dtype=np.float32)[0]
+
+    def search(value=effort):
+        return table.search(vector).distance_type(metric).search_effort(value).limit(8)
+
+    def partitions(query):
+        values = re.findall(r"partitions_searched=(\d+)", query.analyze_plan())
+        assert values
+        return sum(map(int, values))
+
+    assert partitions(search()) == probes
+    exact = set(
+        table.search(vector)
+        .distance_type(metric)
+        .bypass_vector_index()
+        .limit(8)
+        .to_arrow()["id"]
+        .to_pylist()
+    )
+    actual = set(search().to_arrow()["id"].to_pylist())
+    assert len(actual & exact) / 8 >= 0.5
+    expected = 1 if effort is None else probes
+    assert partitions(search().nprobes(1)) == expected
+    assert (
+        partitions(
+            table.search(vector)
+            .distance_type(metric)
+            .nprobes(1)
+            .search_effort(effort)
+            .limit(8)
+        )
+        == expected
+    )
+    assert partitions(search().nprobes(1).search_effort(None)) == 1
+    assert partitions(search().maximum_nprobes(2)) <= 2
+    assert partitions(search().minimum_nprobes(3).maximum_nprobes(3)) == 3
+    assert partitions(search(0).nprobes(16).maximum_nprobes(0)) == 1
+    if effort is None:
+        assert search().to_arrow() == search(0.5).to_arrow()
+    filtered = search(0).maximum_nprobes(2).where("id % 8 = 0", prefilter=True)
+    assert partitions(filtered) <= 2
+    assert all(row["id"] % 8 == 0 for row in filtered.to_list())
+
+
+@pytest.mark.parametrize(
+    "effort", [-0.1, 1.1, float("nan"), float("inf"), -float("inf")]
+)
+def test_search_effort_invalid_sync(table, effort):
+    with pytest.raises(ValueError, match="search_effort must be finite and in"):
+        table.search([0, 0]).search_effort(effort).to_list()
+
+
+@pytest.mark.asyncio
+@pytest.mark.parametrize("hybrid", [False, True])
+@pytest.mark.parametrize("effort", [None, 0, 0.25, 0.5, 0.75, 1])
+async def test_search_effort_async_roundtrip(table_async, hybrid, effort):
+    query = table_async.vector_search([0, 0])
+    if hybrid:
+        query = query.nearest_to_text("dog")
+    query = query.nprobes(5).search_effort(effort).minimum_nprobes(2).maximum_nprobes(3)
+    request = query.to_query_object()
+    assert request.search_effort == effort
+    assert request.nprobes == 5
+    assert request.minimum_nprobes == 2
+    assert request.maximum_nprobes == 3
+    assert query.search_effort(None).to_query_object().search_effort is None
+    query.search_effort(effort)
+    if not hybrid:
+        assert len(await query.limit(1).to_list()) == 1
+
+
+@pytest.mark.asyncio
+@pytest.mark.parametrize("hybrid", [False, True])
+@pytest.mark.parametrize(
+    "effort", [-0.1, 1.1, float("nan"), float("inf"), -float("inf")]
+)
+async def test_search_effort_invalid_async(table_async, hybrid, effort):
+    query = table_async.vector_search([0, 0])
+    if hybrid:
+        query = query.nearest_to_text("dog")
+    with pytest.raises(ValueError, match="search_effort must be finite and in"):
+        query.search_effort(effort)
diff --git a/python/src/query.rs b/python/src/query.rs
index 88b7243d4..e5547881a 100644
--- a/python/src/query.rs
+++ b/python/src/query.rs
@@ -71,7 +71,7 @@ impl<'a, 'py> FromPyObject<'a, 'py> for PyLanceDB<FtsQuery> {
         match class_name(&ob)?.as_str() {
             "MatchQuery" => {
                 let query = ob.getattr("query")?.extract()?;
-                let column = ob.getattr("column")?.extract()?;
+                let column = ob.getattr("column")?.extract::<Option<String>>()?;
                 let boost = ob.getattr("boost")?.extract()?;
                 let fuzziness = ob.getattr("fuzziness")?.extract()?;
                 let max_expansions = ob.getattr("max_expansions")?.extract()?;
@@ -88,7 +88,7 @@ impl<'a, 'py> FromPyObject<'a, 'py> for PyLanceDB<FtsQuery> {
 
                 let mut query =
                     MatchQuery::new(query)
-                        .with_column(Some(column))
+                        .with_column(column)
                         .with_boost(boost)
                         .with_fuzziness(fuzziness)
                         .with_max_expansions(max_expansions)
@@ -199,7 +199,7 @@ impl<'py> IntoPyObject<'py> for PyLanceDB<FtsQuery> {
                 }
                 namespace
                     .getattr(intern!(py, "MatchQuery"))?
-                    .call((query.terms, query.column.unwrap()), Some(&kwargs))
+                    .call((query.terms, query.column), Some(&kwargs))
             }
             FtsQuery::Phrase(query) => {
                 let kwargs = PyDict::new(py);
@@ -339,6 +339,7 @@ pub struct PyQueryRequest {
     pub nprobes: Option<usize>,
     pub minimum_nprobes: Option<usize>,
     pub maximum_nprobes: Option<usize>,
+    pub search_effort: Option<f64>,
     pub lower_bound: Option<f32>,
     pub upper_bound: Option<f32>,
     pub ef: Option<usize>,
@@ -371,6 +372,7 @@ impl From<AnyQuery> for PyQueryRequest {
                 nprobes: None,
                 minimum_nprobes: None,
                 maximum_nprobes: None,
+                search_effort: None,
                 lower_bound: None,
                 upper_bound: None,
                 ef: None,
@@ -399,6 +401,7 @@ impl From<AnyQuery> for PyQueryRequest {
                 nprobes: vector_query.nprobes,
                 minimum_nprobes: vector_query.minimum_nprobes,
                 maximum_nprobes: vector_query.maximum_nprobes,
+                search_effort: vector_query.search_effort,
                 lower_bound: vector_query.lower_bound,
                 upper_bound: vector_query.upper_bound,
                 ef: vector_query.ef,
@@ -998,6 +1001,15 @@ impl VectorQuery {
         Ok(())
     }
 
+    pub fn search_effort(&mut self, search_effort: Option<f64>) -> PyResult<()> {
+        self.inner = self
+            .inner
+            .clone()
+            .search_effort(search_effort)
+            .infer_error()?;
+        Ok(())
+    }
+
     pub fn minimum_nprobes(&mut self, minimum_nprobes: u32) -> PyResult<()> {
         self.inner = self
             .inner
@@ -1191,6 +1203,18 @@ impl HybridQuery {
         self.inner_vec.nprobes(nprobe)
     }
 
+    pub fn search_effort(&mut self, search_effort: Option<f64>) -> PyResult<()> {
+        self.inner_vec.search_effort(search_effort)
+    }
+
+    pub fn minimum_nprobes(&mut self, minimum_nprobes: u32) -> PyResult<()> {
+        self.inner_vec.minimum_nprobes(minimum_nprobes)
+    }
+
+    pub fn maximum_nprobes(&mut self, maximum_nprobes: u32) -> PyResult<()> {
+        self.inner_vec.maximum_nprobes(maximum_nprobes)
+    }
+
     pub fn ef(&mut self, ef: u32) {
         self.inner_vec.ef(ef);
     }
@@ -1235,6 +1259,10 @@ impl HybridQuery {
         req.query_vector = vec_req.query_vector;
         req.column = vec_req.column;
         req.distance_type = vec_req.distance_type;
+        req.nprobes = vec_req.nprobes;
+        req.minimum_nprobes = vec_req.minimum_nprobes;
+        req.maximum_nprobes = vec_req.maximum_nprobes;
+        req.search_effort = vec_req.search_effort;
         req.ef = vec_req.ef;
         req.refine_factor = vec_req.refine_factor;
         req.lower_bound = vec_req.lower_bound;
diff --git a/python/uv.lock b/python/uv.lock
index b013e5982..ace8b98b9 100644
--- a/python/uv.lock
+++ b/python/uv.lock
@@ -1900,6 +1900,7 @@ clip = [
     { name = "torch" },
 ]
 dev = [
+    { name = "maturin" },
     { name = "pre-commit" },
     { name = "pyright" },
     { name = "ruff" },
@@ -1981,6 +1982,7 @@ requires-dist = [
     { name = "ibm-watsonx-ai", marker = "python_full_version >= '3.10' and extra == 'embeddings'", specifier = ">=1.1.2" },
     { name = "instructorembedding", marker = "extra == 'embeddings'", specifier = ">=1.0.1" },
     { name = "lance-namespace", specifier = ">=0.13.0" },
+    { name = "maturin", marker = "extra == 'dev'", specifier = ">=1.10" },
     { name = "mkdocs", marker = "extra == 'docs'" },
     { name = "mkdocs-jupyter", marker = "extra == 'docs'" },
     { name = "mkdocs-material", marker = "extra == 'docs'" },
@@ -2005,7 +2007,7 @@ requires-dist = [
     { name = "pyarrow-stubs", marker = "extra == 'tests'", specifier = ">=16.0" },
     { name = "pydantic", specifier = ">=2.7.4,<3" },
     { name = "pylance", marker = "extra == 'pylance'", specifier = ">=5.0.0b5" },
-    { name = "pylance", marker = "extra == 'tests'", specifier = "==14.0.0b10" },
+    { name = "pylance", marker = "extra == 'tests'", specifier = "==14.0.0b12" },
     { name = "pyright", marker = "extra == 'dev'", specifier = ">=1.1.350" },
     { name = "pytest", marker = "extra == 'tests'", specifier = ">=7.0" },
     { name = "pytest-asyncio", marker = "extra == 'tests'", specifier = ">=0.21" },
@@ -2156,6 +2158,30 @@ wheels = [
     { url = "https://files.pythonhosted.org/packages/41/09/5b161152e2d90f7b87f781c2e1267494aef9c32498df793f73ad0a0a494a/matplotlib_inline-0.2.2-py3-none-any.whl", hash = "sha256:3c821cf1c209f59fb2d2d64abbf5b23b67bcb2210d663f9918dd851c6da1fcf6", size = 9534, upload-time = "2026-05-08T17:33:32.055Z" },
 ]
 
+[[package]]
+name = "maturin"
+version = "1.15.0"
+source = { registry = "https://pypi.org/simple" }
+dependencies = [
+    { name = "tomli", marker = "python_full_version < '3.11'" },
+]
+sdist = { url = "https://files.pythonhosted.org/packages/b9/c8/22e5e21b2679c9bce6415ca578034ca2cc9316be0642ae21e051a2d5198c/maturin-1.15.0.tar.gz", hash = "sha256:94b26cc8e8aba61a5f2099715fe640e18c5f678e9a500408b38761263954228a", size = 385504, upload-time = "2026-08-24T12:11:22.665Z" }
+wheels = [
+    { url = "https://files.pythonhosted.org/packages/14/69/5c01b461044eb1f45ddcce006706eb88110c793cdb11c7ae0b5e08492e94/maturin-1.15.0-py3-none-linux_armv6l.whl", hash = "sha256:6bf6dc62e22d4dcfd5a51244ff0d58975fa4979c48209fe84159617648956d82", size = 10206220, upload-time = "2026-08-24T12:10:53.327Z" },
+    { url = "https://files.pythonhosted.org/packages/eb/1f/2b431554e11687cdb1077e0cdadcc118c53f611086b3af00c8545a67c6a5/maturin-1.15.0-py3-none-macosx_10_12_x86_64.macosx_11_0_arm64.macosx_10_12_universal2.whl", hash = "sha256:cd35772633f489841132bc8e71d6fc7f842df30b9c05cd5cdf1ee1ddcb744cc7", size = 19416513, upload-time = "2026-08-24T12:10:56.126Z" },
+    { url = "https://files.pythonhosted.org/packages/51/36/e23a21cb34a648b711036b9b2fe1d4f3f4ee24f8db54215d73f1a9a3a3ec/maturin-1.15.0-py3-none-macosx_10_12_x86_64.whl", hash = "sha256:c40b4eae7bf5ef1f4b1af8d623fe4105016f93578fb15b764e741d08ec3b92dd", size = 10014962, upload-time = "2026-08-24T12:10:58.486Z" },
+    { url = "https://files.pythonhosted.org/packages/8c/80/33b15cb2d8f30f12c807955e8f2fd775027692904e30ec0784744ce8cd83/maturin-1.15.0-py3-none-manylinux_2_12_i686.manylinux2010_i686.musllinux_1_1_i686.whl", hash = "sha256:7eb066372f541f8eb4909c79c5d9bd0b9e8125980bdf1ec9e8aba23c6c8d6c55", size = 10196223, upload-time = "2026-08-24T12:11:00.696Z" },
+    { url = "https://files.pythonhosted.org/packages/fe/91/b495e19e2f5c503b540452b2039115e7b2363867e8c5ad4179eb752fa92c/maturin-1.15.0-py3-none-manylinux_2_12_x86_64.manylinux2010_x86_64.musllinux_1_1_x86_64.whl", hash = "sha256:653020a63525bb224e5ab0adf02e17a2e08bc86dbea7fc1399c9a56d7529b99e", size = 10541186, upload-time = "2026-08-24T12:11:02.857Z" },
+    { url = "https://files.pythonhosted.org/packages/5d/2b/2abff58037188d852b124871b1f0d720e1c2bfb3d4f1b03d87c52cd66488/maturin-1.15.0-py3-none-manylinux_2_17_aarch64.manylinux2014_aarch64.musllinux_1_1_aarch64.whl", hash = "sha256:0ebf9767892725083138e671c34482c660317a2f3d6a29fc0e0f34e9d8c99136", size = 10083468, upload-time = "2026-08-24T12:11:05.012Z" },
+    { url = "https://files.pythonhosted.org/packages/3f/07/b7e9f8be99a6627849e81ac7b6694876bce8f50a92995fe17e3cf2610f0a/maturin-1.15.0-py3-none-manylinux_2_17_armv7l.manylinux2014_armv7l.musllinux_1_1_armv7l.whl", hash = "sha256:7ab7eebffd7b8debca2265985de4eaeb332141276d24b9560b5ad484d4b3add1", size = 10047786, upload-time = "2026-08-24T12:11:07.042Z" },
+    { url = "https://files.pythonhosted.org/packages/ca/ab/167e3cb7accee11b507dbe53e0e87aeccb376d44ae66284c96ee4df3a9fd/maturin-1.15.0-py3-none-manylinux_2_17_ppc64le.manylinux2014_ppc64le.musllinux_1_1_ppc64le.whl", hash = "sha256:126e12e618b4db42f68c779a56d41f82a390145ba36ac3f621d057eb34f5ad9d", size = 13315332, upload-time = "2026-08-24T12:11:09.433Z" },
+    { url = "https://files.pythonhosted.org/packages/14/4d/801379f646cbc6b00998e5289b0630a886be3a4ee4c75b6bc9b87478a7f1/maturin-1.15.0-py3-none-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:4f9d33e6c3f9615c8caceecbbbd440f8eb25a3ddeb687077682cd5eca2e9ae15", size = 10807183, upload-time = "2026-08-24T12:11:11.73Z" },
+    { url = "https://files.pythonhosted.org/packages/89/27/2e612e1cbd1580e9e94d4722c227b5180dca27b32b955a34b79918aa1292/maturin-1.15.0-py3-none-manylinux_2_31_riscv64.musllinux_1_1_riscv64.whl", hash = "sha256:bf29beddd0c6708f112db51d5275fc28b28b9e9c9c5faae387eaef662918b176", size = 10413274, upload-time = "2026-08-24T12:11:14.04Z" },
+    { url = "https://files.pythonhosted.org/packages/70/d8/202a7b4d75a51f20f84ec9ce3b7345b12b822207072164e2b1c6ef665125/maturin-1.15.0-py3-none-win32.whl", hash = "sha256:da649988be98e87e009e51b1bf0d301b6a301bc0cecbdd60d40d8ba60748d1ca", size = 8928744, upload-time = "2026-08-24T12:11:16.269Z" },
+    { url = "https://files.pythonhosted.org/packages/40/dc/4e90da594986ba78dd3bc8a5921ecdcdb11085b22b02a412caab3b225601/maturin-1.15.0-py3-none-win_amd64.whl", hash = "sha256:552c2be4afd43fe8d5c9f3ec8d4c4756d973b8dcbe94c14084390301f50243e1", size = 10335085, upload-time = "2026-08-24T12:11:18.326Z" },
+    { url = "https://files.pythonhosted.org/packages/8b/10/15d4314edf130955edf2dc237aa393a8a7c10f2b9b57b89fa2f61f915659/maturin-1.15.0-py3-none-win_arm64.whl", hash = "sha256:c7dc0c66c78d3debdd9c5aa807e861fbcbf07f3505d34b125df74c03986b0f48", size = 9713795, upload-time = "2026-08-24T12:11:20.83Z" },
+]
+
 [[package]]
 name = "mdit-py-plugins"
 version = "0.6.1"
@@ -3912,7 +3938,7 @@ crypto = [
 
 [[package]]
 name = "pylance"
-version = "14.0.0b10"
+version = "14.0.0b12"
 source = { registry = "https://pypi.fury.io/lance-format/" }
 dependencies = [
     { name = "lance-namespace" },
@@ -3922,12 +3948,12 @@ dependencies = [
     { name = "pyarrow" },
 ]
 wheels = [
-    { url = "https://pypi.fury.io/lance-format/-/ver_4hvE/pylance-14.0.0b10-cp310-abi3-macosx_11_0_arm64.whl", hash = "sha256:415b6b1b9fe735fbae7c755eb9774ae2fe0feddc90bf80bee24087cbfbfdf8c1" },
-    { url = "https://pypi.fury.io/lance-format/-/ver_pM1km/pylance-14.0.0b10-cp310-abi3-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:87317cad22f4d35fea11eb70aa5c0d71a0e1ea3b96de38f3deca933daa749fff" },
-    { url = "https://pypi.fury.io/lance-format/-/ver_24jdyj/pylance-14.0.0b10-cp310-abi3-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:6da554a2fc2197ba254676b12824d8120b48d65937697e0661a61302458a4740" },
-    { url = "https://pypi.fury.io/lance-format/-/ver_XclaZ/pylance-14.0.0b10-cp310-abi3-manylinux_2_28_aarch64.whl", hash = "sha256:36c80a801675bb9dcbd681fa99d0cdd5ca642dfb3ca08d71ac980278a2f7748b" },
-    { url = "https://pypi.fury.io/lance-format/-/ver_L0yh/pylance-14.0.0b10-cp310-abi3-manylinux_2_28_x86_64.whl", hash = "sha256:e50ad0b866faefb2fa3ae87328b93fa0d06b2bdd5bb5d0c1a6c74b907cc2fd93" },
-    { url = "https://pypi.fury.io/lance-format/-/ver_2gbNr8/pylance-14.0.0b10-cp310-abi3-win_amd64.whl", hash = "sha256:3d86451fe00f71510005bb700737fb1a4f9361554541cbf802f3c5072554d014" },
+    { url = "https://pypi.fury.io/lance-format/-/ver_1Ixasc/pylance-14.0.0b12-cp310-abi3-macosx_11_0_arm64.whl", hash = "sha256:84350a3c5c1eea3398f4c4fad7cfe85b831d831a5e43b5ca2a783dbdb0e11969" },
+    { url = "https://pypi.fury.io/lance-format/-/ver_sCEkG/pylance-14.0.0b12-cp310-abi3-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:76b386c182a6a8e7b86fbef28f84d1dd39e4626ac8681a0cf36f40a45dd7cb87" },
+    { url = "https://pypi.fury.io/lance-format/-/ver_1vdT1C/pylance-14.0.0b12-cp310-abi3-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:6aa548c7937fcdfa8f1943a2d59041a42f8b500199880dbb509c7195fa253381" },
+    { url = "https://pypi.fury.io/lance-format/-/ver_2fXu1V/pylance-14.0.0b12-cp310-abi3-manylinux_2_28_aarch64.whl", hash = "sha256:79f82d6a6c50f37aab338470e3733436e6a27fbb5972d29fcf453ece56d3b3c6" },
+    { url = "https://pypi.fury.io/lance-format/-/ver_24UavN/pylance-14.0.0b12-cp310-abi3-manylinux_2_28_x86_64.whl", hash = "sha256:de0305999257002785f60169e6926896584d66f1da050bb5afd1d48c6e3d9bb9" },
+    { url = "https://pypi.fury.io/lance-format/-/ver_XNynX/pylance-14.0.0b12-cp310-abi3-win_amd64.whl", hash = "sha256:4d0cdb689e898ce71e8e434d00a3b97639cea8fb97da21f1c921543be363268f" },
 ]
 
 [[package]]
diff --git a/rust/lancedb/src/query.rs b/rust/lancedb/src/query.rs
index 8fb56ceca..07db69bb3 100644
--- a/rust/lancedb/src/query.rs
+++ b/rust/lancedb/src/query.rs
@@ -1127,6 +1127,11 @@ pub struct VectorQueryRequest {
     /// If unset, Lance's default is used. Zero represents an explicitly unbounded
     /// maximum on request and wire representations.
     pub maximum_nprobes: Option<usize>,
+    /// Optional initial IVF search effort in [0, 1].
+    ///
+    /// None preserves Auto and fixed `nprobes`; any explicit value overrides
+    /// `nprobes` while retaining explicit minimum and maximum bounds.
+    pub search_effort: Option<f64>,
     /// The lower bound (inclusive) of the distance to search for.
     pub lower_bound: Option<f32>,
     /// The upper bound (exclusive) of the distance to search for.
@@ -1153,6 +1158,7 @@ impl Default for VectorQueryRequest {
             nprobes: None,
             minimum_nprobes: None,
             maximum_nprobes: None,
+            search_effort: None,
             lower_bound: None,
             upper_bound: None,
             ef: None,
@@ -1165,6 +1171,17 @@ impl Default for VectorQueryRequest {
 }
 
 impl VectorQueryRequest {
+    pub(crate) fn validate_search_effort(&self) -> Result<()> {
+        if let Some(effort) = self.search_effort
+            && (!effort.is_finite() || !(0.0..=1.0).contains(&effort))
+        {
+            return Err(Error::InvalidInput {
+                message: format!("search_effort must be finite and in [0, 1], got {effort}"),
+            });
+        }
+        Ok(())
+    }
+
     fn add_query_vector(&mut self, vector: impl IntoQueryVector) -> Result<()> {
         let vector = vector.to_query_vector(&DataType::Float32, "default")?;
         let array =
@@ -1279,7 +1296,8 @@ impl VectorQuery {
     /// you the desired recall.
     ///
     /// LanceDB retains this as `nprobes` through local and remote request
-    /// construction. Lance sets both probe bounds to this value. For more
+    /// construction. Lance sets both probe bounds to this value only when
+    /// [`Self::search_effort`] is unset. For more
     /// fine-grained control see [`VectorQuery::minimum_nprobes`] and
     /// [`VectorQuery::maximum_nprobes`].
     ///
@@ -1294,6 +1312,31 @@ impl VectorQuery {
         Ok(self)
     }
 
+    /// Set the initial IVF search effort to a finite value in [0, 1].
+    ///
+    /// None (the default) resolves to 0.5, preserving Auto probing. At 0 the
+    /// search starts at the caller minimum; at 1 it searches all partitions
+    /// within the caller maximum. Intermediate values interpolate around Auto.
+    /// Later probing may continue when too few candidates are found.
+    /// Any explicit value, including 0.5, overrides [`Self::nprobes`] regardless
+    /// of setter order. Explicit minimum and maximum bounds still apply.
+    /// This does not change refinement or HNSW ef and is not a recall guarantee.
+    /// Lance's MemWAL LSM scanner does not yet support this option.
+    ///
+    /// ```
+    /// # use lancedb::{Table, query::{QueryBase, ExecutableQuery}};
+    /// # async fn example(table: &Table) -> lancedb::Result<()> {
+    /// let results = table.query().nearest_to([0.1, 0.2].as_slice())?
+    ///     .search_effort(0.75)?.execute().await?;
+    /// # Ok(())
+    /// # }
+    /// ```
+    pub fn search_effort(mut self, search_effort: impl Into<Option<f64>>) -> Result<Self> {
+        self.request.search_effort = search_effort.into();
+        self.request.validate_search_effort()?;
+        Ok(self)
+    }
+
     /// Set the minimum number of partitions to search
     ///
     /// This argument is only used when the vector column has an IVF PQ index.
@@ -1333,6 +1376,8 @@ impl VectorQuery {
     ///
     /// If this value is greater than minimum_nprobes then the excess partitions will
     /// only be searched if the initial search does not return enough results.
+    /// The bound applies per index segment to initial and later probing without
+    /// disabling Auto probing.
     ///
     /// This can be useful when there is a narrow filter to allow these queries to
     /// spend more time searching and avoid potential false negatives.
@@ -2448,6 +2493,26 @@ mod tests {
         assert_eq!(query.request.nprobes, None);
         assert_eq!(query.request.minimum_nprobes, None);
         assert_eq!(query.request.maximum_nprobes, None);
+        assert_eq!(query.request.search_effort, None);
+
+        let effort_query = query
+            .clone()
+            .nprobes(20)
+            .unwrap()
+            .search_effort(0.5)
+            .unwrap();
+        assert_eq!(effort_query.request.search_effort, Some(0.5));
+        assert_eq!(effort_query.request.nprobes, Some(20));
+        assert_eq!(effort_query.request.minimum_nprobes, None);
+        assert_eq!(effort_query.request.maximum_nprobes, None);
+        assert_eq!(
+            effort_query
+                .search_effort(None)
+                .unwrap()
+                .request
+                .search_effort,
+            None
+        );
         assert_eq!(
             *query
                 .request
diff --git a/rust/lancedb/src/remote/table.rs b/rust/lancedb/src/remote/table.rs
index 119577dc5..356fb1e1a 100644
--- a/rust/lancedb/src/remote/table.rs
+++ b/rust/lancedb/src/remote/table.rs
@@ -1137,6 +1137,7 @@ impl<S: HttpSend> RemoteTable<S> {
         mut body: serde_json::Value,
         query: &VectorQueryRequest,
     ) -> Result<Vec<serde_json::Value>> {
+        query.validate_search_effort()?;
         self.apply_query_params(&mut body, &query.base)?;
 
         // Apply general parameters, before we dispatch based on number of query vectors.
@@ -1155,6 +1156,9 @@ impl<S: HttpSend> RemoteTable<S> {
         if let Some(maximum_nprobes) = query.maximum_nprobes {
             body["maximum_nprobes"] = maximum_nprobes.into();
         }
+        if let Some(search_effort) = query.search_effort {
+            body["search_effort"] = search_effort.into();
+        }
         body["lower_bound"] = query.lower_bound.into();
         body["upper_bound"] = query.upper_bound.into();
         body["ef"] = query.ef.into();
@@ -6269,6 +6273,77 @@ mod tests {
         ));
     }
 
+    #[rstest]
+    #[case::unset(None)]
+    #[case::minimum(Some(0.0))]
+    #[case::low(Some(0.25))]
+    #[case::auto(Some(0.5))]
+    #[case::high(Some(0.75))]
+    #[case::maximum(Some(1.0))]
+    #[tokio::test]
+    async fn test_query_vector_search_effort_wire(#[case] effort: Option<f64>) {
+        let table = Table::new_with_handler("my_table", move |request| {
+            let body = request_body_json(&request);
+            assert_eq!(body["nprobes"], 7);
+            assert_eq!(body["minimum_nprobes"], 2);
+            assert_eq!(body["maximum_nprobes"], 8);
+            if let Some(effort) = effort {
+                assert_eq!(body["search_effort"], effort);
+            } else {
+                assert!(body.get("search_effort").is_none());
+            }
+            let batch = RecordBatch::try_new(
+                Arc::new(Schema::new(vec![Field::new("id", DataType::Int32, false)])),
+                vec![Arc::new(Int32Array::from(vec![1]))],
+            )
+            .unwrap();
+            http::Response::builder()
+                .status(200)
+                .header(CONTENT_TYPE, ARROW_FILE_CONTENT_TYPE)
+                .body(write_ipc_file(&batch))
+                .unwrap()
+        });
+        let batches = table
+            .query()
+            .nearest_to(vec![0.1, 0.2])
+            .unwrap()
+            .search_effort(effort)
+            .unwrap()
+            .nprobes(7)
+            .unwrap()
+            .minimum_nprobes(2)
+            .unwrap()
+            .maximum_nprobes(Some(8))
+            .unwrap()
+            .execute()
+            .await
+            .unwrap()
+            .try_collect::<Vec<_>>()
+            .await
+            .unwrap();
+        assert_eq!(batches[0].num_rows(), 1);
+    }
+
+    #[rstest]
+    #[case::negative(-0.1)]
+    #[case::too_large(1.1)]
+    #[case::nan(f64::NAN)]
+    #[case::infinity(f64::INFINITY)]
+    #[case::negative_infinity(f64::NEG_INFINITY)]
+    #[tokio::test]
+    async fn test_query_vector_search_effort_invalid(#[case] effort: f64) {
+        let table = Table::new_with_handler::<&str>("my_table", |_| {
+            panic!("invalid search_effort must not send a request")
+        });
+        let result = table
+            .query()
+            .nearest_to(vec![0.1, 0.2])
+            .unwrap()
+            .search_effort(effort);
+        assert!(matches!(result, Err(Error::InvalidInput { message })
+            if message.contains("search_effort must be finite and in [0, 1]")));
+    }
+
     #[tokio::test]
     async fn test_query_vector_all_params() {
         let table = Table::new_with_handler("my_table", |request| {
diff --git a/rust/lancedb/src/table/query.rs b/rust/lancedb/src/table/query.rs
index 77c4ca1a5..3ab54ae07 100644
--- a/rust/lancedb/src/table/query.rs
+++ b/rust/lancedb/src/table/query.rs
@@ -106,7 +106,7 @@ async fn can_execute_namespace_query(table: &NativeTable, query: &AnyQuery) -> R
 }
 
 fn requires_local_namespace_execution(query: &AnyQuery) -> bool {
-    // The namespace QueryTable request cannot represent approx_mode or use_lsm.
+    // The namespace QueryTable request cannot represent approx_mode, search_effort or use_lsm.
     // For use_lsm that is worse than a tuning miss: MemWAL read routing lives only
     // in `create_plan`, so a pushed-down query would return stale base-only data.
     if query.base().use_lsm.is_some() || query.base().take_offsets.is_some() {
@@ -115,7 +115,7 @@ fn requires_local_namespace_execution(query: &AnyQuery) -> bool {
     matches!(
         query,
         AnyQuery::VectorQuery(vector_query)
-            if vector_query.approx_mode.is_some()
+            if vector_query.approx_mode.is_some() || vector_query.search_effort.is_some()
     )
 }
 
@@ -163,6 +163,7 @@ pub async fn create_plan(
         AnyQuery::Query(query) => VectorQueryRequest::from_plain_query(query),
     };
     query.base.check_filter()?;
+    query.validate_search_effort()?;
 
     let ds_ref = table.dataset.get().await?;
 
@@ -293,7 +294,9 @@ pub async fn create_plan(
             scanner.approx_mode(approx_mode.into());
         }
 
-        if let Some(nprobes) = query.nprobes {
+        if query.search_effort.is_none()
+            && let Some(nprobes) = query.nprobes
+        {
             if query.maximum_nprobes == Some(0) {
                 scanner.minimum_nprobes(nprobes);
             } else {
@@ -306,6 +309,9 @@ pub async fn create_plan(
         if let Some(maximum_nprobes) = query.maximum_nprobes.filter(|nprobes| *nprobes != 0) {
             scanner.maximum_nprobes(maximum_nprobes);
         }
+        if let Some(search_effort) = query.search_effort {
+            scanner.search_effort(search_effort)?;
+        }
     }
 
     // For a batch query, `nearest` already applies k to each query vector.
@@ -604,6 +610,11 @@ fn convert_to_namespace_query(query: &AnyQuery) -> Result<NsQueryTableRequest> {
             };
 
             // Check for unsupported features
+            if vq.search_effort.is_some() {
+                return Err(Error::NotSupported {
+                    message: "namespace QueryTable does not support search_effort; execute the query locally".to_string(),
+                });
+            }
             if vq.base.reranker.is_some() {
                 return Err(Error::NotSupported {
                     message: "Reranker is not supported for server-side queries".to_string(),
@@ -962,6 +973,19 @@ mod tests {
         let default_query = AnyQuery::VectorQuery(VectorQueryRequest::default());
         assert!(!requires_local_namespace_execution(&default_query));
 
+        for effort in [0.0, 0.5, 1.0] {
+            let effort_query = AnyQuery::VectorQuery(VectorQueryRequest {
+                query_vector: vec![Arc::new(Float32Array::from(vec![1.0, 2.0]))],
+                search_effort: Some(effort),
+                ..Default::default()
+            });
+            assert!(requires_local_namespace_execution(&effort_query));
+            assert!(matches!(
+                convert_to_namespace_query(&effort_query),
+                Err(Error::NotSupported { message }) if message.contains("search_effort")
+            ));
+        }
+
         let nprobes_query = AnyQuery::VectorQuery(VectorQueryRequest {
             query_vector: vec![Arc::new(Float32Array::from(vec![1.0, 2.0]))],
             nprobes: Some(20),
diff --git a/rust/lancedb/src/table/query/lsm.rs b/rust/lancedb/src/table/query/lsm.rs
index 698b9c736..20e36e95d 100644
--- a/rust/lancedb/src/table/query/lsm.rs
+++ b/rust/lancedb/src/table/query/lsm.rs
@@ -174,6 +174,9 @@ fn reject_unsupported(query: &VectorQueryRequest) -> Result<()> {
     // brute-force search the index-only base arm can't do. (ef / approx_mode are
     // recall/speed knobs and are left to no-op.)
     if !query.query_vector.is_empty() {
+        if query.search_effort.is_some() {
+            return unsupported("search_effort on vector search");
+        }
         if query.lower_bound.is_some() || query.upper_bound.is_some() {
             return unsupported("distance_range on vector search");
         }
@@ -821,6 +824,19 @@ mod tests {
     use super::*;
     use lance_index::mem_wal::{CompactedSsTable, IndexCatchupProgress};
 
+    #[test]
+    fn search_effort_is_not_silently_ignored() {
+        let query = VectorQueryRequest {
+            query_vector: vec![Arc::new(arrow_array::Float32Array::from(vec![1.0, 0.0]))],
+            search_effort: Some(0.5),
+            ..Default::default()
+        };
+        assert!(
+            matches!(reject_unsupported(&query), Err(Error::NotSupported { message })
+            if message.contains("search_effort") && message.contains("un-compacted MemWAL"))
+        );
+    }
+
     #[test]
     fn resident_index_checks_distinguish_columns_and_query_types() {
         use arrow_schema::Field;