feat: Add custom metadata ingestion and filtering

alembic/versions/0008_metadata_fields.py

@@ -0,0 +1,39 @@
+"""custom metadata field catalog
+
+Revision ID: 0008_metadata_fields
+Revises: 0007_add_knowledge_delete_anonymous
+Create Date: 2026-07-16 00:00:00.000000
+"""
+
+from collections.abc import Sequence
+
+import sqlalchemy as sa
+
+from alembic import op
+
+revision: str = "0008_metadata_fields"
+down_revision: str | Sequence[str] | None = "0007_add_knowledge_delete_anonymous"
+branch_labels: str | Sequence[str] | None = None
+depends_on: str | Sequence[str] | None = None
+
+
+def upgrade() -> None:
+    op.create_table(
+        "metadata_fields",
+        sa.Column("key", sa.String(length=64), nullable=False),
+        sa.Column("metadata_type", sa.String(length=16), nullable=False),
+        sa.Column("created_at", sa.DateTime(), nullable=False),
+        sa.Column("updated_at", sa.DateTime(), nullable=False),
+        sa.PrimaryKeyConstraint("key", name="pk_metadata_fields"),
+    )
+    op.create_index(
+        "ix_metadata_fields_metadata_type",
+        "metadata_fields",
+        ["metadata_type"],
+        unique=False,
+    )
+
+
+def downgrade() -> None:
+    op.drop_index("ix_metadata_fields_metadata_type", table_name="metadata_fields")
+    op.drop_table("metadata_fields")

frontend/app/api/queries/useGetSearchQuery.ts

@@ -3,6 +3,7 @@ import {
   useQuery,
   useQueryClient,
 } from "@tanstack/react-query";
+import type { MetadataGroup } from "@/components/metadata-filter-builder";
 import type { ParsedQueryData } from "@/contexts/knowledge-filter-context";
 import { SEARCH_CONSTANTS } from "@/lib/constants";
 import { buildSearchPayloadFilters } from "@/lib/filter-normalization";
@@ -16,6 +17,7 @@ export interface SearchPayload {
     document_types?: string[];
     owners?: string[];
     connector_types?: string[];
+    metadata?: MetadataGroup;
   };
 }
 
@@ -39,6 +41,7 @@ export interface ChunkResult {
   index?: number;
   allowed_users?: string[];
   allowed_groups?: string[];
+  metadata?: Record<string, { type: string; value: unknown }>;
 }
 
 export interface File {

frontend/app/chat/_types/types.ts

@@ -88,6 +88,7 @@ export interface SelectedFilters {
   document_types: string[];
   owners: string[];
   connector_types: string[];
+  metadata?: import("@/components/metadata-filter-builder").MetadataGroup;
 }
 
 export interface KnowledgeFilterData {
@@ -109,6 +110,7 @@ export interface RequestBody {
     document_types?: string[];
     owners?: string[];
     connector_types?: string[];
+    metadata?: import("@/components/metadata-filter-builder").MetadataGroup;
   };
   filter_id?: string;
   limit?: number;

frontend/components/cloud-picker/custom-metadata-editor.tsx

@@ -0,0 +1,177 @@
+"use client";
+
+import { Plus, Trash2 } from "lucide-react";
+import { useEffect, useMemo, useState } from "react";
+import { Button } from "@/components/ui/button";
+import { Input } from "@/components/ui/input";
+import {
+  Select,
+  SelectContent,
+  SelectItem,
+  SelectTrigger,
+  SelectValue,
+} from "@/components/ui/select";
+import type { CustomMetadataEntry, CustomMetadataType } from "./types";
+
+interface MetadataFieldSuggestion {
+  key: string;
+  type: CustomMetadataType;
+}
+
+interface Props {
+  value: CustomMetadataEntry[];
+  onChange: (value: CustomMetadataEntry[]) => void;
+}
+
+const defaultValue = (type: CustomMetadataType): string | number | boolean => {
+  if (type === "number") return 0;
+  if (type === "boolean") return false;
+  return "";
+};
+
+export function CustomMetadataEditor({ value, onChange }: Props) {
+  const [suggestions, setSuggestions] = useState<MetadataFieldSuggestion[]>([]);
+
+  useEffect(() => {
+    let active = true;
+    fetch("/api/metadata/fields")
+      .then((response) => (response.ok ? response.json() : { fields: [] }))
+      .then((result) => {
+        if (active) setSuggestions(result.fields ?? []);
+      })
+      .catch(() => undefined);
+    return () => {
+      active = false;
+    };
+  }, []);
+
+  const fieldTypes = useMemo(
+    () => new Map(suggestions.map((field) => [field.key, field.type])),
+    [suggestions],
+  );
+
+  const update = (index: number, patch: Partial<CustomMetadataEntry>) => {
+    const next = [...value];
+    next[index] = { ...next[index], ...patch };
+    onChange(next);
+  };
+
+  return (
+    <div className="mt-6 space-y-3 border-t pt-4">
+      <div>
+        <div className="text-sm font-semibold">Custom metadata</div>
+        <div className="text-sm text-muted-foreground">
+          Applied to every document selected for this ingest.
+        </div>
+      </div>
+      <datalist id="custom-metadata-keys">
+        {suggestions.map((field) => (
+          <option key={field.key} value={field.key} />
+        ))}
+      </datalist>
+      {value.map((entry, index) => {
+        const knownType = fieldTypes.get(entry.key.trim().toLowerCase());
+        return (
+          <div
+            className="grid grid-cols-[minmax(0,1fr)_130px_minmax(0,1fr)_36px] gap-2"
+            key={`${index}-${entry.key}`}
+          >
+            <Input
+              aria-label="Metadata key"
+              list="custom-metadata-keys"
+              placeholder="supplier"
+              value={entry.key}
+              onChange={(event) => {
+                const key = event.target.value.toLowerCase();
+                const suggestedType = fieldTypes.get(key);
+                update(index, {
+                  key,
+                  ...(suggestedType && suggestedType !== entry.type
+                    ? {
+                        type: suggestedType,
+                        value: defaultValue(suggestedType),
+                      }
+                    : {}),
+                });
+              }}
+            />
+            <Select
+              disabled={!!knownType}
+              value={knownType ?? entry.type}
+              onValueChange={(type: CustomMetadataType) =>
+                update(index, { type, value: defaultValue(type) })
+              }
+            >
+              <SelectTrigger aria-label="Metadata type">
+                <SelectValue />
+              </SelectTrigger>
+              <SelectContent>
+                <SelectItem value="string">Text</SelectItem>
+                <SelectItem value="number">Number</SelectItem>
+                <SelectItem value="date">Date</SelectItem>
+                <SelectItem value="boolean">Boolean</SelectItem>
+              </SelectContent>
+            </Select>
+            {entry.type === "boolean" ? (
+              <Select
+                value={String(entry.value)}
+                onValueChange={(next) =>
+                  update(index, { value: next === "true" })
+                }
+              >
+                <SelectTrigger aria-label="Metadata value">
+                  <SelectValue />
+                </SelectTrigger>
+                <SelectContent>
+                  <SelectItem value="true">True</SelectItem>
+                  <SelectItem value="false">False</SelectItem>
+                </SelectContent>
+              </Select>
+            ) : (
+              <Input
+                aria-label="Metadata value"
+                type={
+                  entry.type === "date"
+                    ? "date"
+                    : entry.type === "number"
+                      ? "number"
+                      : "text"
+                }
+                value={String(entry.value)}
+                onChange={(event) =>
+                  update(index, {
+                    value:
+                      entry.type === "number"
+                        ? Number(event.target.value)
+                        : event.target.value,
+                  })
+                }
+              />
+            )}
+            <Button
+              aria-label="Remove metadata"
+              onClick={() =>
+                onChange(value.filter((_, itemIndex) => itemIndex !== index))
+              }
+              size="icon"
+              type="button"
+              variant="ghost"
+            >
+              <Trash2 className="h-4 w-4" />
+            </Button>
+          </div>
+        );
+      })}
+      <Button
+        onClick={() =>
+          onChange([...value, { key: "", type: "string", value: "" }])
+        }
+        size="sm"
+        type="button"
+        variant="outline"
+      >
+        <Plus className="h-4 w-4" /> Add metadata
+      </Button>
+    </div>
+  );
+}

frontend/components/cloud-picker/ingest-settings.tsx

@@ -35,6 +35,7 @@ import {
 } from "@/components/ui/tooltip";
 import { useAuth } from "@/contexts/auth-context";
 import { knowledgeToIngestSettings } from "@/lib/ingest-settings-knowledge";
+import { CustomMetadataEditor } from "./custom-metadata-editor";
 import type { IngestSettings as IngestSettingsType } from "./types";
 
 interface IngestSettingsProps {
@@ -285,6 +286,12 @@ export const IngestSettings = ({
             </div>
           </div>
         )}
+        {showAdvancedSettings && (
+          <CustomMetadataEditor
+            value={currentSettings.metadata ?? []}
+            onChange={(metadata) => handleSettingsChange({ metadata })}
+          />
+        )}
 
         <div className={showAdvancedSettings ? "" : "mt-6"}>
           {showShared && (

frontend/components/cloud-picker/types.ts

@@ -115,6 +115,17 @@ export interface IngestSettings {
   embeddingModel: string;
   /** When true, index without an owner so all users in the instance can retrieve the document. COS only. */
   shared?: boolean;
+  /** Typed custom metadata applied to every document in the ingest batch. */
+  metadata?: CustomMetadataEntry[];
+}
+
+export type CustomMetadataType = "string" | "number" | "date" | "boolean";
+export type CustomMetadataValue = string | number | boolean;
+
+export interface CustomMetadataEntry {
+  key: string;
+  type: CustomMetadataType;
+  value: CustomMetadataValue;
 }
 
 /** Inline error message if chunk settings are invalid; otherwise null. */