Google does not publish a Google Lens API. Its own public API directory lists 314 APIs including Cloud Vision and Document AI, with no Lens entry at all, so what gets sold under that name is either a search scraper priced per query or Google Cloud OCR priced per 1,000 units. Upload a receipt or invoice below to see the fields extracted and priced per page.
Accepts PDF, JPG, PNG, BMP, HEIC and TIFF
Upload your receipts and invoices
Drop files here or click to upload
Upload failed, please try again
Up to 50 files
Uploading...
Every other page on this query answers a question Google never asked. There is no Lens product to buy, so the rate cards people quote belong to other products entirely, and two of those products are not even in the same category as each other. Pricing the work properly means first deciding which of the two jobs you are actually doing.
Google publishes a machine readable directory of every public API it offers. When fetched, it returned 314 distinct APIs. Cloud Vision is there as vision and Document AI is there as documentai. There is no lens entry, and there is no Lens page in the Google Cloud catalog or in the API Explorer.
A Google Lens API can mean a scraper that replays Lens visual search results, which finds where an image appears and what product it matches, or it can mean Google Cloud OCR, which reads the text printed on an image. They cost different amounts because they do unrelated things, and buyers routinely price one while planning to build the other.
One unit is one feature applied to one image. Running Text Detection and Label Detection on the same photograph bills two units, not one. Any estimate built on an image count rather than a feature count is low by exactly the number of features you plan to call.
Enterprise Document OCR runs $1.50 per 1,000 pages. The Expense parser that returns merchant, date, tax and total as named fields is $0.10 per document, which is $100 per 1,000. Both are Document AI, both are on the same price list, and the gap between them is the entire question of who writes the parsing.
Scrapingdog charges 5 API credits for one Lens request, so a plan advertised as 200,000 credits is 40,000 Lens calls. Comparing that headline against a competitor quoted in searches rather than credits gets the ranking wrong by a factor of five before you start.
Cloud Vision gives 1,000 free units a month and Document AI gives 1,000 free pages. Those are generous for a prototype and irrelevant to a real workload. The first month you process 5,000 receipts you pay for 4,000 of them, and the per unit rate you skimmed past is suddenly the whole budget.
Most people who search for a Google Lens API want the thing Lens appears to do on a phone: point it at a document and get the useful numbers back. Neither the scrapers nor the raw OCR endpoints do that. If the documents are receipts, invoices or expense paperwork, upload a batch below and check the columns that come back.
Vendor, date, invoice number, subtotal, sales tax, total and every line item come back as their own columns. Raw text detection returns the characters and their coordinates and leaves the meaning to you.
Output lands in a spreadsheet a bookkeeper can work from directly, or as structured JSON your code can read, without a formatting step in between.
The Starter plan is $49 a month for 2,500 pages, about two cents a page. There is no per feature multiplier, so asking for line items as well as header fields does not bill twice.
Drop a folder in and get a file back. No billing account to enable, no service account key to rotate, no quota request to file before a busy month.
The same extraction is available as a REST call returning the same fields as JSON, so a prototype in the browser and the production integration use one schema.
A published monthly rate with an annual option and a free tier you can try without a card. No quote request stands between you and a number.
Work through these in order. Most bad estimates in this category come from skipping the first step and pricing a product that does not do the job.
Visual search answers where else does this image appear and what product is this. Text extraction answers what does this document say. If your input is a receipt, an invoice or a form, you want text extraction and the Lens scrapers are the wrong aisle entirely.
Tip: A quick test: if a plain photocopy of the document would still satisfy your user, you want OCR. If the answer depends on recognizing the object itself, you want visual search.
Cloud Vision bills one unit per feature per image. List the features you will call on each image, multiply by your monthly image count, and that product is your unit count. Two features on 10,000 images is 20,000 units, not 10,000.
Raw OCR at $1.50 per 1,000 pages is cheap because it stops at characters. A parser at $0.10 per document is more expensive because it returns named fields. The difference is engineering time, and at low volume that time costs far more than the rate gap ever will.
Subtract the 1,000 free units, price the remainder at the published rate, then add the tier below it if you expect to cross 5,000,000 units. Compare that annual figure against a flat per page plan before you commit engineering effort to either.
Every rate below was read from the vendor: Google Cloud from its own pricing pages, SerpApi and Scrapingdog from theirs, and the API directory from the Google discovery endpoint. Where a figure is our arithmetic on those published rates rather than a vendor rate, it says so.
You were asked to add Google Lens to a product and need to explain what that actually costs before the sprint is planned.
You want the fields off a document, and you are trying to work out whether a cloud parser or a per page tool is cheaper at your volume.
You genuinely want Lens style image matching and need to know which unofficial providers exist and what they charge.
You are putting Google, AWS and Azure side by side and need every rate expressed in one comparable unit.
No. Google does not publish a Google Lens API. Google maintains a machine readable directory of its public APIs, and a fetch of that directory returned 314 distinct APIs. Cloud Vision appears as vision and Document AI appears as documentai. There is no lens entry. Every product marketed as a Google Lens API is either an unofficial scraper of Lens search results or a rebadged reference to Google Cloud Vision.
There is no official rate because there is no official product. The three real ways to pay for the work people mean are Cloud Vision at $1.50 per 1,000 units after a free 1,000, Document AI parsers at $0.10 per document, and third party Lens scrapers from roughly $0.45 to $25 per 1,000 requests depending on provider and plan. Which of those applies depends entirely on whether you want text or visual matches.
Cloud Vision is the product Google itself points developers toward when they ask for Lens. Prices are per 1,000 units, the first 1,000 units each month are free, and one unit is one feature applied to one image. Applying two features to the same image bills two units.
| Feature | First 1,000 units/month | Units 1,001 to 5,000,000 | Above 5,000,000 |
|---|---|---|---|
| Text Detection | Free | $1.50 | $0.60 |
| Document Text Detection | Free | $1.50 | $0.60 |
| Label Detection | Free | $1.50 | $1.00 |
| Web Detection | Free | $3.50 | Contact Google |
| Safe Search Detection | Free | Free with Label Detection, or $1.50 | Free with Label Detection, or $0.60 |
Web Detection is the closest thing Google sells to the reverse image lookup that Lens performs in a phone camera, and at $3.50 per 1,000 it is the most expensive line on the card. If what you admired about Lens was that it identifies an object rather than reads it, this row is your real starting price, not the text rows.
Document AI is the other half of Google OCR and it is where the named fields live. The gap between the raw OCR line and the parser lines is the single most important number on this page: $1.50 per 1,000 pages against $100 per 1,000 documents for the same paperwork.
| Processor | Published rate | Per 1,000 documents | What you get back |
|---|---|---|---|
| Enterprise Document OCR | $1.50 per 1,000 pages, first 1,000 free | $1.50 | Text and layout, no meaning |
| Expense parser | $0.10 per document | $100.00 | Receipt fields as named entities |
| Invoice parser | $0.10 per document | $100.00 | Invoice header and line items |
| Custom extractor | $30.00 per 1,000, then $20.00 above 1,000,000 | $30.00 | Your own trained schema |
| OCR add ons | $6.00 per 1,000 | $6.00 | Extra OCR capabilities on top |
The Expense parser is described on the price list as the former receipt parser, so if you have older notes quoting a receipt parser rate, this is the same line renamed. Google also publishes discounted rates under one year and three year commitment plans, which bring OCR to $1.35 and $1.20 per 1,000 and the parsers to $0.09 and $0.08 per document. Those require opting into a consumption model, so treat the standard column as your price unless procurement has already signed one.
These providers scrape and structure Lens visual search results. They do not read documents, and none of them is affiliated with Google. Rates below are the published plans; the per 1,000 column is our arithmetic.
| Provider | Entry paid plan | Included volume | Per 1,000 Lens requests | Free tier |
|---|---|---|---|---|
| SerpApi | Starter, $25 per month | 1,000 searches | $25.00 | 250 searches per month |
| SerpApi Production | $150 per month | 15,000 searches | $10.00 | Same |
| Scrapingdog | Lite, $40 per month | 200,000 credits at 5 per Lens request | $1.00 | 200 credits |
| Scrapingdog Standard | $90 per month | 1,000,000 credits | $0.45 | Same |
| Apify | Usage based platform credit | Not published on the Lens actor page | Not published | $5 of platform credit monthly |
Read the Scrapingdog row carefully, because the credit weighting is where comparisons in this category go wrong. One Lens request costs 5 credits, so the 200,000 credit Lite plan is 40,000 Lens calls rather than 200,000. We have left the Apify per request figure blank rather than estimate it, because the Lens actor page publishes a free credit allowance and no per result rate, and a number we cannot reconcile to the vendor does not belong in a pricing table.
Here is one realistic workload, 5,000 receipts a month, priced under each approach. Google figures use the published rates above, ReceiptOCR figures are our own published plan rates, and the monthly totals are arithmetic.
| Approach | Monthly cost at 5,000 receipts | Parsing work you still owe |
|---|---|---|
| Cloud Vision Text Detection | About $6.00 | All of it. You receive text and coordinates. |
| Document AI Enterprise OCR | About $6.00 | All of it, with better layout to work from. |
| Document AI Expense parser | $500.00 | None. Named fields come back. |
| ReceiptOCR Plus | $149.00 for 10,000 pages | None. Named fields and line items. |
| A Lens scraper | Not applicable | Wrong product. Returns visual matches. |
The two cheap rows are genuinely cheap and they are the right answer for some projects. If you have engineers, predictable layouts and time, $6 a month for raw OCR plus your own parsing beats every managed option on cash cost. The reason most teams do not end there is that receipt layouts are not predictable, and the parser you write to handle 200 merchant formats becomes something somebody has to maintain every quarter. Price the engineering, not just the API call.
Yes, Google Lens reads text from images, and that is why developers look for its API. The text recognition inside Lens is built on the same underlying Google vision technology exposed through Cloud Vision. The difference is packaging: Lens is a consumer feature inside Search, Photos and Android, while Cloud Vision is the billable developer product. You can use Lens yourself on a phone at no cost, but you cannot call it from your own software.
It depends on which half of the job you need. For raw text at the lowest possible rate, Cloud Vision Document Text Detection at $1.50 per 1,000 units is hard to beat. For named fields with no parsing work, the Document AI Expense and Invoice parsers at $0.10 per document are the Google answer, and a per page extractor is the cheaper one at volume. For visual search rather than text, SerpApi and Scrapingdog are the established unofficial providers.
No. Google does not publish a Google Lens API. Its public API discovery directory returned 314 APIs, including Cloud Vision as vision and Document AI as documentai, with no lens entry. Anything sold as a Google Lens API is an unofficial scraper of Lens results or a reference to Cloud Vision.
There is no official price because there is no official product. The real options are Cloud Vision at $1.50 per 1,000 units after 1,000 free, Document AI parsers at $0.10 per document, and third party Lens scrapers from about $0.45 to $25 per 1,000 requests. Which applies depends on whether you need text or visual matches.
Google Lens is a consumer feature inside Search, Photos and Android that you use by hand and cannot call from code. Cloud Vision is the billable developer API that exposes comparable text and image analysis. Lens is free to use and unavailable to developers; Cloud Vision is available to developers and metered per feature.
Yes. Google Lens reads printed and handwritten text from images and is built on the same vision technology Google exposes through Cloud Vision. The practical limit is that Lens is an end user feature, so you can point your own phone at a document but you cannot send images to Lens from your own application.
Google Lens is free to use as an app feature, and there is no API to price either way. The closest official product, Cloud Vision, gives 1,000 free units per month and bills $1.50 per 1,000 after that. Document AI also gives 1,000 free pages a month on its OCR processor.
You cannot call Lens directly. The supported route is Cloud Vision for text and label detection or Document AI for structured document fields. On Android, ML Kit provides on device text recognition at no per call cost, which is the closest thing to Lens behavior inside your own application.
It charges per feature per image, not per image. One unit is one feature applied to one image, so running Text Detection and Label Detection on the same photo bills two units at $1.50 per 1,000 each. Budget on your feature count multiplied by image count, never on image count alone.
Because it returns meaning rather than characters. Enterprise Document OCR is $1.50 per 1,000 pages and gives you text and layout. The Expense and Invoice parsers are $0.10 per document, or $100 per 1,000, and give back named fields like merchant, date, tax and total. You are paying for the parsing you would otherwise write.
Raw Cloud Vision at about $6 for 5,000 receipts is cheapest on cash cost but leaves you all the parsing. Document AI Expense parser costs $500 for the same 5,000 and needs none. A per page extractor sits between them: ReceiptOCR covers 10,000 pages for $149 a month with fields and line items included.
SerpApi, Scrapingdog, Bright Data and Apify are established commercial companies with published rates, but none is affiliated with or endorsed by Google, and all of them work by scraping Lens results rather than by licensed access. Treat availability as something that can change, and read the terms before building a business on one.
Five API credits per request, which its Google Lens documentation states directly. That makes the $40 Lite plan with 200,000 credits worth 40,000 Lens calls, about $1.00 per 1,000 requests, and the $90 Standard plan with 1,000,000 credits worth 200,000 calls at about $0.45 per 1,000. Those per 1,000 figures are our arithmetic.
Not as a file you can use. Lens can copy text off a receipt to your clipboard on a phone, one document at a time, with no fields and no export. Getting vendor, date, sales tax, total and line items into Excel or CSV across a batch needs a document extractor, which is what this page exists to price.
Stop typing receipts by hand
Upload your receipts and invoices and get a clean Excel or CSV file in minutes.
Extract my receipts nowFree to try, no sign up required
Named fields instead of raw text, for the same documents.
The full processor rate card behind the parser lines above.
Every published OCR and extraction meter side by side, per 1,000 pages.
The same extracted fields as structured JSON, priced per page.
What Lens style text recognition does with handwritten documents.
Where each engine fits once you know what you are extracting.
Get started extracting receipt and invoice data to spreadsheets.
USD
per month
per month
$288 charged today
Choose speed vs accuracy when extracting
| Base AI Faster | 2,500 pages |
| Pro AI Best accuracy | 500 pages |
Scale receipt and invoice extraction across your team with automation.
USD
per month
per month
$888 charged today
Choose speed vs accuracy when extracting
| Base AI Faster | 10,000 pages |
| Pro AI Best accuracy | 2,000 pages |
Enterprise‑grade receipt and invoice extraction and controls.
USD
per month
per month
$ charged today
Choose speed vs accuracy when extracting
| Base AI Faster | pages |
| Pro AI Best accuracy | pages |
Custom workflows and deployment for AI automation at scale