Merge branch 'master' into a11y/announce-streamed-reply

Author: timothycarambat

README.md

@@ -188,9 +188,9 @@ Mintplex Labs & the community maintain a number of deployment methods, scripts,
 |----------------------------------------|----|-----|---------------|------------|
 | [![Deploy on Docker][docker-btn]][docker-deploy] | [![Deploy on AWS][aws-btn]][aws-deploy] | [![Deploy on GCP][gcp-btn]][gcp-deploy] | [![Deploy on DigitalOcean][do-btn]][do-deploy] | [![Deploy on Render.com][render-btn]][render-deploy] |
 
-| Railway                                             | RepoCloud                                                 | Elestio                                             | Northflank                                                   |
-| --------------------------------------------------- | --------------------------------------------------------- | --------------------------------------------------- | ------------------------------------------------------------ |
-| [![Deploy on Railway][railway-btn]][railway-deploy] | [![Deploy on RepoCloud][repocloud-btn]][repocloud-deploy] | [![Deploy on Elestio][elestio-btn]][elestio-deploy] | [![Deploy on Northflank][northflank-btn]][northflank-deploy] |
+| Railway                                             | RepoCloud                                                 | Elestio                                             | Northflank                                                   | Sealos                                               |
+| --------------------------------------------------- | --------------------------------------------------------- | --------------------------------------------------- | ------------------------------------------------------------ | ---------------------------------------------------- |
+| [![Deploy on Railway][railway-btn]][railway-deploy] | [![Deploy on RepoCloud][repocloud-btn]][repocloud-deploy] | [![Deploy on Elestio][elestio-btn]][elestio-deploy] | [![Deploy on Northflank][northflank-btn]][northflank-deploy] | [![Deploy on Sealos][sealos-btn]][sealos-deploy] |
 
 [or set up a production AnythingLLM instance without Docker →](./BARE_METAL.md)
 
@@ -316,3 +316,5 @@ This project is [MIT](./LICENSE) licensed.
 [elestio-deploy]: https://elest.io/open-source/anythingllm
 [northflank-btn]: https://assets.northflank.com/deploy_to_northflank_smm_36700fb050.svg
 [northflank-deploy]: https://northflank.com/stacks/deploy-anythingllm
+[sealos-btn]: https://sealos.io/Deploy-on-Sealos.svg
+[sealos-deploy]: https://sealos.io/products/app-store/anything-llm

SECURITY.md

@@ -46,6 +46,12 @@ If your report is about being able to access the system via _bypassing the authe
 
 If your report depends on an attacker knowing or guessing a UUID (e.g., a session ID, embed ID, or any other random identifier) without demonstrating a practical way to obtain it, this is not a valid report. UUIDs (v4) have 122 bits of entropy and are not feasible to brute-force. Unless your report includes a concrete method to leak or enumerate the UUID in question, it will be closed immediately.
 
+### Reports based on Developer API Key access
+
+Developer API keys are **system-level credentials**. They are only creatable by administrators and intentionally grant full, unrestricted access to the entire `/v1/*` API surface — equivalent to admin access. If your report's attack chain starts with "attacker has a developer API key," you are describing an attacker who already has admin-equivalent access. There is no privilege escalation possible from a developer API key because it is already the highest privilege level.
+
+This includes reports about API keys being able to mint auth tokens, access admin endpoints, or perform actions on behalf of other users. All of these are intended capabilities of the API key system. If you believe a developer API key should have scoped or limited permissions, that is a feature request, not a vulnerability.
+
 ### Reports about admin-enabled agent tools doing "too much"
 
 Several agent tools (e.g., `sql-agent`, filesystem tools) are disabled by default and require an administrator to explicitly enable and configure them. If your report is that an admin-enabled tool can perform actions beyond what you think it should (writes via SQL, broad file access, etc.), this is not a valid report. The admin who enables the tool and configures its access (connection strings, paths, credentials) is making an intentional decision about what the agent can do. We provide UI-level warnings but intentionally do not restrict functionality — the user should be able to do whatever they want with their own systems.

collector/__tests__/utils/OCRLoader/index.test.js

@@ -0,0 +1,69 @@
+/* eslint-env jest, node */
+const OCRLoader = require("../../../utils/OCRLoader");
+
+describe("toSharpRawInput", () => {
+  test("expands packed 1-bit DeviceGray so Sharp gets integer channels", () => {
+    const packed = Buffer.from([0b10101010]);
+    const raw = OCRLoader.toSharpRawInput({
+      width: 8,
+      height: 1,
+      data: packed,
+      kind: 1,
+    });
+    expect(raw.channels).toBe(1);
+    expect(raw.data).toEqual(Buffer.from([255, 0, 255, 0, 255, 0, 255, 0]));
+  });
+
+  test("handles widths that are not a multiple of 8 (row padding bits)", () => {
+    // width 10 -> rowBytes = 2, last 6 bits of each row are padding.
+    // Row 0: 1111111111 (padding 000000), row 1: 0000000000 (padding 111111)
+    const packed = Buffer.from([0b11111111, 0b11000000, 0b00000000, 0b00111111]);
+    const raw = OCRLoader.toSharpRawInput({
+      width: 10,
+      height: 2,
+      data: packed,
+      kind: 1,
+    });
+    expect(raw.channels).toBe(1);
+    expect(raw.data.length).toBe(20);
+    expect(raw.data).toEqual(
+      Buffer.from([...Array(10).fill(255), ...Array(10).fill(0)])
+    );
+  });
+
+  test("detects packed 1-bit by buffer length when kind is omitted", () => {
+    const width = 2480;
+    const height = 2;
+    const packed = Buffer.alloc(Math.ceil(width / 8) * height, 0xff);
+    const raw = OCRLoader.toSharpRawInput({ width, height, data: packed });
+    expect(raw.channels).toBe(1);
+    expect(raw.data.length).toBe(width * height);
+    expect(raw.data[0]).toBe(255);
+  });
+
+  test("passes through 8-bit gray and copies the buffer", () => {
+    const data = Buffer.from([1, 2, 3, 4]);
+    const raw = OCRLoader.toSharpRawInput({ width: 2, height: 2, data });
+    expect(raw.channels).toBe(1);
+    data[0] = 99;
+    expect(raw.data[0]).toBe(1);
+  });
+
+  test("passes through RGB", () => {
+    const data = Buffer.alloc(2 * 2 * 3, 128);
+    const raw = OCRLoader.toSharpRawInput({ width: 2, height: 2, data });
+    expect(raw.channels).toBe(3);
+    expect(raw.data.length).toBe(12);
+  });
+
+  test("returns null for empty or mismatched buffers", () => {
+    expect(OCRLoader.toSharpRawInput(null)).toBeNull();
+    expect(
+      OCRLoader.toSharpRawInput({
+        width: 8,
+        height: 1,
+        data: Buffer.from([1, 2]),
+      })
+    ).toBeNull();
+  });
+});

collector/utils/OCRLoader/index.js

@@ -274,6 +274,67 @@ class OCRLoader {
   }
 }
 
+/**
+ * pdf.js ImageKind.GRAYSCALE_1BPP. Packed 1-bit rows are ceil(width/8) bytes.
+ * @type {number}
+ */
+const PDFJS_GRAYSCALE_1BPP = 1;
+
+/**
+ * Unpack pdf.js packed 1-bit DeviceGray into 8-bit grayscale Sharp can ingest.
+ * Bit 1 is white, bit 0 is black (PDF DeviceGray).
+ *
+ * @param {Buffer|Uint8Array} packed
+ * @param {number} width
+ * @param {number} height
+ * @returns {Buffer|null}
+ */
+function unpackPackedGray1Bpp(packed, width, height) {
+  if (!packed || width <= 0 || height <= 0) return null;
+  const rowBytes = Math.ceil(width / 8);
+  if (packed.length < rowBytes * height) return null;
+
+  const out = Buffer.alloc(width * height);
+  for (let y = 0; y < height; y++) {
+    const rowStart = y * rowBytes;
+    for (let x = 0; x < width; x++) {
+      const byte = packed[rowStart + (x >> 3)];
+      const bit = (byte >> (7 - (x & 7))) & 1;
+      out[y * width + x] = bit ? 255 : 0;
+    }
+  }
+  return out;
+}
+
+/**
+ * Build a Sharp `raw` input from a pdf.js image object.
+ * Copies the source buffer so concurrent OCR workers cannot share backing memory.
+ * 1-bit CCITT/bitonal scans (issue #6118) are expanded to 8-bit gray.
+ *
+ * @param {{width?: number, height?: number, data?: ArrayLike<number>, kind?: number}} img
+ * @returns {{data: Buffer, width: number, height: number, channels: number}|null}
+ */
+function toSharpRawInput(img) {
+  const width = img?.width;
+  const height = img?.height;
+  if (!width || !height || !img.data) return null;
+
+  const src = Buffer.from(img.data);
+  const isPacked1Bpp =
+    img.kind === PDFJS_GRAYSCALE_1BPP ||
+    src.length === Math.ceil(width / 8) * height;
+
+  if (isPacked1Bpp) {
+    const unpacked = unpackPackedGray1Bpp(src, width, height);
+    if (!unpacked) return null;
+    return { data: unpacked, width, height, channels: 1 };
+  }
+
+  const channels = src.length / width / height;
+  if (![1, 2, 3, 4].includes(channels)) return null;
+  return { data: src, width, height, channels };
+}
+
 /**
  * Converts a PDF page to a buffer using Sharp.
  * @param {Object} options - The options for the Sharp PDF page object.
@@ -313,14 +374,14 @@ class PDFSharp {
 
           const name = ops.argsArray[i][0];
           const img = await page.objs.get(name);
-          const { width, height } = img;
-          const size = img.data.length;
-          const channels = size / width / height;
+          const raw = toSharpRawInput(img);
+          if (!raw) continue;
+          const { data, width, height, channels } = raw;
           const targetDPI = 70;
           const targetWidth = Math.floor(width * (targetDPI / 72));
           const targetHeight = Math.floor(height * (targetDPI / 72));
 
-          const image = this.sharp(img.data, {
+          const image = this.sharp(data, {
             raw: { width, height, channels },
             density: targetDPI,
           })
@@ -352,4 +413,6 @@ class PDFSharp {
   }
 }
 
+OCRLoader.toSharpRawInput = toSharpRawInput;
+OCRLoader.unpackPackedGray1Bpp = unpackPackedGray1Bpp;
 module.exports = OCRLoader;

frontend/public/anything-llm-dark.png

@@ -6,6 +6,10 @@ const DEFAULT_MODELS = [
     id: "gemini-embedding-001",
     name: "Gemini Embedding 001",
   },
+  {
+    id: "gemini-embedding-2",
+    name: "Gemini Embedding 2",
+  },
 ];
 
 export default function GeminiOptions({ settings }) {

frontend/public/anything-llm-light.png

@@ -128,35 +128,45 @@ export default function AwsBedrockLLMOptions({ settings }) {
   );
 }
 
+const MANUAL_MODEL_ENTRY = "-- Enter model ID manually --";
+
 function BedrockModelSelection({ settings, apiKey, region }) {
   const [groupedModels, setGroupedModels] = useState({});
   const [loading, setLoading] = useState(true);
+  const [manualEntry, setManualEntry] = useState(false);
 
   useEffect(() => {
     async function findCustomModels() {
       setLoading(true);
-      const { models } = await System.customModels(
+      const { models = [] } = await System.customModels(
         "bedrock",
         apiKey,
         null,
         null,
         { region }
       );
-      if (models?.length > 0) {
-        const modelsByOrganization = models.reduce((acc, model) => {
-          const org = model.organization || "AWS Bedrock";
-          acc[org] = acc[org] || [];
-          acc[org].push(model);
-          return acc;
-        }, {});
-        setGroupedModels(modelsByOrganization);
-      }
+      const modelsByOrganization = models.reduce((acc, model) => {
+        const org = model.organization || "AWS Bedrock";
+        acc[org] = acc[org] || [];
+        acc[org].push(model);
+        return acc;
+      }, {});
+      setGroupedModels(modelsByOrganization);
+
+      // Saved models not present in the fetched list (eg: cross-region
+      // inference profile IDs the Mantle listing omits) can only render
+      // via manual entry - same for an empty list.
+      const savedModel = settings?.AwsBedrockLLMModel;
+      const savedModelInList = models.some((model) => model.id === savedModel);
+      setManualEntry(
+        models.length === 0 || (!!savedModel && !savedModelInList)
+      );
       setLoading(false);
     }
     findCustomModels();
   }, [apiKey, region]);
 
-  if (loading || Object.keys(groupedModels).length === 0) {
+  if (loading) {
     return (
       <div className="flex flex-col w-60">
         <label className="text-white text-sm font-semibold block mb-3">
@@ -175,6 +185,35 @@ function BedrockModelSelection({ settings, apiKey, region }) {
     );
   }
 
+  if (manualEntry) {
+    return (
+      <div className="flex flex-col w-60">
+        <label className="text-white text-sm font-semibold block mb-3">
+          Chat Model Selection
+        </label>
+        <input
+          type="text"
+          name="AwsBedrockLLMModel"
+          className="border-none bg-theme-settings-input-bg text-white placeholder:text-theme-settings-input-placeholder text-sm rounded-lg focus:outline-primary-button active:outline-primary-button outline-none block w-full p-2.5"
+          placeholder="eu.anthropic.claude-sonnet-4-5-20250929-v1:0"
+          defaultValue={settings?.AwsBedrockLLMModel}
+          required={true}
+          autoComplete="off"
+          spellCheck={false}
+        />
+        {Object.keys(groupedModels).length > 0 && (
+          <button
+            type="button"
+            onClick={() => setManualEntry(false)}
+            className="text-white/60 hover:text-white text-xs text-left mt-1.5 underline w-fit"
+          >
+            Select from available models
+          </button>
+        )}
+      </div>
+    );
+  }
+
   return (
     <div className="flex flex-col w-60">
       <label className="text-white text-sm font-semibold block mb-3">
@@ -183,6 +222,9 @@ function BedrockModelSelection({ settings, apiKey, region }) {
       <select
         name="AwsBedrockLLMModel"
         required={true}
+        onChange={(e) => {
+          if (e.target.value === MANUAL_MODEL_ENTRY) setManualEntry(true);
+        }}
         className="border-none bg-theme-settings-input-bg border-gray-500 text-white text-sm rounded-lg block w-full p-2.5"
       >
         {Object.keys(groupedModels)
@@ -200,6 +242,8 @@ function BedrockModelSelection({ settings, apiKey, region }) {
               ))}
             </optgroup>
           ))}
+        <option disabled={true}>──────────</option>
+        <option value={MANUAL_MODEL_ENTRY}>{MANUAL_MODEL_ENTRY}</option>
       </select>
     </div>
   );

frontend/src/components/EmbeddingSelection/GeminiOptions/index.jsx

@@ -140,7 +140,7 @@ export default function SettingsSidebar() {
       <div>
         <Link
           to={paths.home()}
-          className="flex shrink-0 max-w-[55%] items-center justify-start mx-[20.5px] my-[18px]"
+          className="flex shrink-0 items-center justify-start mx-[20.5px] my-[18px]"
         >
           <img
             src={logo}