// Get index info to check last mutation processed
const info = await env.VECTORIZE_INDEX.describe();
console.log(info.mutationId); // Last mutation ID
console.log(info.processedUpToMutation); // Last processed timestamp

# 1. Create the index with FIXED dimensions and metric
npx wrangler vectorize create my-index \
  --dimensions=768 \
  --metric=cosine

# 2. Create metadata indexes IMMEDIATELY (before inserting vectors!)
npx wrangler vectorize create-metadata-index my-index \
  --property-name=category \
  --type=string

npx wrangler vectorize create-metadata-index my-index \
  --property-name=timestamp \
  --type=number

# Dimensions MUST match your embedding model output:
# - Workers AI @cf/baai/bge-base-en-v1.5: 768 dimensions
# - OpenAI text-embedding-3-small: 1536 dimensions
# - OpenAI text-embedding-3-large: 3072 dimensions

# Metrics determine similarity calculation:
# - cosine: Best for normalized embeddings (most common)
# - euclidean: Absolute distance between vectors
# - dot-product: For non-normalized vectors

{
  "name": "my-vectorize-worker",
  "main": "src/index.ts",
  "compatibility_date": "2025-10-21",
  "vectorize": [
    {
      "binding": "VECTORIZE_INDEX",
      "index_name": "my-index"
    }
  ],
  "ai": {
    "binding": "AI"
  }
}

export interface Env {
  VECTORIZE_INDEX: VectorizeIndex;
  AI: Ai;
}

interface VectorizeVector {
  id: string;
  values: number[] | Float32Array | Float64Array;
  namespace?: string;
  metadata?: Record<string, string | number | boolean | string[]>;
}

interface VectorizeMatches {
  matches: Array<{
    id: string;
    score: number;
    values?: number[];
    metadata?: Record<string, any>;
    namespace?: string;
  }>;
  count: number;
}

// Equality (implicit $eq)
{ category: "docs" }

// Not equals
{ status: { $ne: "archived" } }

// In/Not in arrays
{ category: { $in: ["docs", "tutorials"] } }
{ category: { $nin: ["deprecated", "draft"] } }

// Range queries (numbers) - NEW in V2
{ timestamp: { $gte: 1704067200, $lt: 1735689600 } }

// Range queries (strings) - prefix searching
{ url: { $gte: "/docs/workers", $lt: "/docs/workersz" } }

// Nested metadata with dot notation
{ "author.id": "user123" }

// Multiple conditions (implicit AND)
{ category: "docs", language: "en", "metadata.published": true }

// Few unique values - efficient filtering
metadata: {
  category: "docs",        // ~10 categories
  language: "en",          // ~5 languages
  published: true          // 2 values (boolean)
}

// Many unique values - avoid large range scans
metadata: {
  user_id: "uuid-v4...",         // Millions of unique values
  timestamp_ms: 1704067200123    // Use seconds instead
}

// ❌ INVALID metadata keys
metadata: {
  "": "value",              // Empty key
  "user.name": "John",      // Contains dot (reserved for nesting)
  "$admin": true,           // Starts with $
  "key\"with\"quotes": 1    // Contains quotes
}

// ✅ VALID metadata keys
metadata: {
  "user_name": "John",
  "isAdmin": true,
  "nested": { "allowed": true }  // Access as "nested.allowed" in filters
}

const BATCH_SIZE = 5000;

async function insertVectors(vectors: VectorizeVector[]) {
  for (let i = 0; i < vectors.length; i += BATCH_SIZE) {
    const batch = vectors.slice(i, i + BATCH_SIZE);
    const result = await env.VECTORIZE.insert(batch);
    console.log(`Inserted batch ${i / BATCH_SIZE + 1}, mutationId: ${result.mutationId}`);

    // Optional: Rate limiting delay
    if (i + BATCH_SIZE < vectors.length) {
      await new Promise(resolve => setTimeout(resolve, 100));
    }
  }
}

// Fast, ~80% accuracy, topK up to 100
const results = await env.VECTORIZE.query(embedding, {
  topK: 50,
  returnValues: false  // Default
});

// Slower, ~100% accuracy, topK max 20
const preciseResults = await env.VECTORIZE.query(embedding, {
  topK: 10,
  returnValues: true   // High-precision scoring
});

Problem: Filtering doesn't work on existing vectors
Solution: Delete and re-insert vectors OR create metadata indexes BEFORE inserting

Problem: "Vector dimensions do not match index configuration"
Solution: Ensure embedding model output matches index dimensions:
  - Workers AI bge-base: 768
  - OpenAI small: 1536
  - OpenAI large: 3072

Problem: "Invalid metadata key"
Solution: Keys cannot:
  - Be empty
  - Contain . (dot)
  - Contain " (quote)
  - Start with $ (dollar sign)

Problem: "Filter exceeds 2048 bytes"
Solution: Simplify filter or split into multiple queries

Problem: Slow queries or reduced accuracy
Solution: Use lower cardinality fields for range queries, or use seconds instead of milliseconds for timestamps

Problem: Updates not reflecting in index
Solution: Use upsert() to overwrite existing vectors, not insert()

Problem: "VECTORIZE_INDEX is not defined"
Solution: Add [[vectorize]] binding to wrangler.jsonc

Problem: Unclear when to use namespace vs metadata filtering
Solution:
  - Namespace: Partition key, applied BEFORE metadata filters
  - Metadata: Flexible key-value filtering within namespace

Problem: Inserted vectors not immediately queryable
Solution: V2 mutations are asynchronous - vectors may take a few seconds to be reflected
  - Use mutationId to track mutation status
  - Check env.VECTORIZE_INDEX.describe() for processedUpToMutation timestamp

Problem: "returnMetadata must be 'all', 'indexed', or 'none'"
Solution: V2 changed returnMetadata from boolean to string enum:
  - ❌ V1: { returnMetadata: true }
  - ✅ V2: { returnMetadata: 'all' }

$ wrangler vectorize list --json
📋 Listing Vectorize indexes...
[
  { "created_on": "2025-10-18T13:28:30.259277Z", ... }
]

# Using tail
wrangler vectorize list --json | tail -n +2 | jq '.'

# Using sed
wrangler vectorize list --json | sed '1d' | jq '.'

const vectorizeRes = env.VECTORIZE.queryById(imgId, {
  filter: { gender: { $in: genderFilters } }, // ❌ TS error but works!
  topK,
  returnMetadata: 'indexed',
});

// Add to your types file
type VectorizeMetadataFilter = Record<string,
  | string
  | number
  | boolean
  | {
      $eq?: string | number | boolean;
      $ne?: string | number | boolean;
      $in?: (string | number | boolean)[];
      $nin?: (string | number | boolean)[];
      $lt?: number | string;
      $lte?: number | string;
      $gt?: number | string;
      $gte?: number | string;
    }
>;

Error: ENOENT: ... '__WRANGLER_EXTERNAL_VECTORIZE_WORKER:<project>:<binding>'

npm install -g wrangler@latest

// ❌ ERROR - topK too high with returnValues
query(embedding, {
  topK: 100,            // Exceeds limit!
  returnValues: true    // Max topK=20 when true
});

// ✅ OK - respects conditional limit
query(embedding, {
  topK: 20,
  returnValues: true
});

// ✅ OK - higher topK without values
query(embedding, {
  topK: 100,
  returnValues: false,
  returnMetadata: 'indexed'
});

// wrangler-test.jsonc (no Vectorize binding)
{
  "name": "my-worker-test",
  "main": "src/index.ts",
  "compatibility_date": "2025-10-21"
  // No vectorize binding
}

// vitest.config.ts
import { defineWorkersProject } from '@cloudflare/vitest-pool-workers/config';

export default defineWorkersProject({
  test: {
    poolOptions: {
      workers: {
        wrangler: {
          configPath: "./wrangler-test.jsonc"
        }
      }
    }
  }
});

// Mock in tests
import { vi } from 'vitest';

const mockVectorize = {
  query: vi.fn().mockResolvedValue({
    matches: [
      { id: 'test-1', score: 0.95, metadata: { category: 'docs' } }
    ],
    count: 1
  }),
  insert: vi.fn().mockResolvedValue({ mutationId: "test-mutation-id" }),
  upsert: vi.fn().mockResolvedValue({ mutationId: "test-mutation-id" })
};

// Use mock in tests
test('vector search', async () => {
  const env = { VECTORIZE_INDEX: mockVectorize };
  // ... test logic
});

// ❌ High-cardinality range at scale
metadata: {
  timestamp_ms: 1704067200123
}
filter: { timestamp_ms: { $gte: 1704067200000 } }

// ✅ Bucketed into discrete values
metadata: {
  timestamp_bucket: "2025-01-01-00:00",  // 1-hour buckets
  timestamp_ms: 1704067200123  // Original (non-indexed)
}
filter: {
  timestamp_bucket: {
    $in: ["2025-01-01-00:00", "2025-01-01-01:00"]
  }
}

const result = await env.VECTORIZE_INDEX.list({
  limit: 1000,  // Max 1000 per page
  cursor?: string
});

// result.vectors: Array<{ id: string }>
// result.cursor: string | undefined
// result.count: number

// Pagination example
let cursor: string | undefined;
const allVectorIds: string[] = [];

do {
  const result = await env.VECTORIZE_INDEX.list({
    limit: 1000,
    cursor
  });
  allVectorIds.push(...result.vectors.map(v => v.id));
  cursor = result.cursor;
} while (cursor);