The following limits apply based on your Workers plan:
| Limit | Workers Free | Workers Paid |
|---|---|---|
| AI Search instances per account | 100 | 5,000 |
| Namespaces per account | 100 | 100 |
| Files per instance | 100,000 | 1M or 500K for hybrid search |
Pages per crawl, discover parse type |
100,000 | 100,000 |
| Maximum PDF file size with OCR enabled | 10 MiB | 10 MiB |
| Maximum plain-text or code file size | 10 MiB | 10 MiB |
| Maximum size for PDFs without OCR and other file formats | 4 MiB | 4 MiB |
| Included semantic queries per month | 1,000 | 1,000 |
| Included full-text queries per month | 1,000 | 1,000 |
| Instances per cross-instance search request | 10 | 10 |
| Maximum pages crawled per day | 500 | Unlimited |
| Max custom metadata fields | 5 per AI Search instance | 5 per AI Search instance |
| Metadata per vector | 10 KiB total, including system overhead | 10 KiB total, including system overhead |
| Filterable indexed string data | First 64 UTF-8 bytes per string | First 64 UTF-8 bytes per string |
Website crawling is bounded by several of these limits at once. A discover crawl accepts up to 100,000 pages, but the files per instance and maximum pages crawled per day limits also apply, so the number of pages you end up with is whichever of those values is lowest. On Workers Free, the daily limit of 500 pages is the binding one.
For the limits that apply only to website data sources, refer to Website.
AI Search billing begins on November 1, 2026. Cloudflare will send a reminder email the week before billing begins.
Every account receives the following included usage each month:
- 5 million ingestion tokens
- 10 GB-month of storage
- 1,000 semantic queries
- 1,000 full-text queries
Usage beyond the included amounts is billed at these rates:
| Usage | Price |
|---|---|
| Base ingestion | $0.75 per million tokens |
| Image processing | An additional $0.50 per million tokens |
| Storage | $2.00 per GB-month |
| Semantic, vector, and hybrid queries | $0.75 per 1,000 queries |
| Full-text queries | $0.10 per 1,000 queries |
Image processing tokens share the 5 million token ingestion allotment. Optical character recognition (OCR) is available on every account. OCR usage is billed as image processing ingestion tokens.
AI Search counts ingestion tokens the same way for every instance, regardless of the embedding model you choose:
- Tokens are counted on the final chunks that AI Search indexes, after parsing and chunking. Because chunk overlap repeats text between neighboring chunks, overlapping text is counted in each chunk it appears in.
- Each chunk is tokenized with the
cl100k_basetokenizer. You can estimate usage for your own content with anycl100k_baseimplementation, such as tiktoken ↗︎. - For images and scanned documents processed with OCR, the extracted text is also counted as image processing tokens, in addition to base ingestion.
Workers AI embedding and reranking usage is included in AI Search usage and does not appear on your Workers AI bill or in your AI Gateway logs. Generation, query rewriting, and external model providers continue to use your account and gateway. Their usage is billed and logged through the applicable service.
Storage, vector indexing, and Browser Run usage for website crawling are included with AI Search. They are not billed separately.
If your instance crawled a website, those pages now live in built-in storage. The dedicated R2 bucket AI Search originally created in your account is no longer used. Objects left in it may still count toward R2 storage usage. AI Search no longer writes to this bucket, so you can delete it if you no longer need its contents.