LlamaParse charges $1.25 per 1,000 credits. Parsing uses 1 to 45 credits a page and field extraction 8 to 60, so one receipt costs from a tenth of a cent to 7.5 cents depending on the tier.
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...
LlamaParse is a capable document parser with a generous seat policy and a clear credit price. The catch for anyone budgeting receipts or invoices is that the credit price is fixed while the credits per page are not. The plan price tells you almost nothing until you know which tier your documents actually need.
The Fast tier costs 1 credit a page, and the LlamaIndex docs say it outputs spatial text only, with no Markdown. Faded thermal receipts and multi-column invoices usually push teams to Agentic at 10 credits or Agentic Plus at 45.
In the v2 API, extract credits per page are the extract tier plus the parse tier. The default Agentic extraction is 15 plus 10, or 25 credits a page, which is $31.25 per 1,000 pages before any retries.
The v1 Invoice preset is listed at 90 credits a page, or $112.50 per 1,000 pages. Integrations written against the older API can be paying several times the v2 rate without anyone noticing.
Parse returns Markdown, JSON or tables. Extract fills a schema you design. Validating totals, splitting sales tax from tips and getting rows into Excel or QuickBooks is work your team builds and maintains.
Most US bookkeepers and AP teams are not shopping for a parsing API. They want vendor, date, sales tax, total and line items in a spreadsheet, checked, with no schema to maintain. ReceiptOCR is that finished workflow on a flat monthly plan.
Vendor, date, invoice number, subtotal, sales tax, tip, total and every line item arrive as columns. There is no extraction schema to design, version or re-test when a vendor changes its layout.
Download one file with consistent headers across the whole batch, ready for a pivot table, a CPA or an import into your accounting system.
Check each extracted value beside the original receipt in the browser before anything is exported, the control a raw API response does not include.
A page is a page. There is no tier choice that multiplies the cost of a hard document, and no parse charge stacked under an extract charge.
Developers can call the same extraction over REST and get parsed receipt and invoice fields as JSON, so a team can start in the browser and automate later.
Starter is $49 a month, or $24 a month billed annually, for 2,500 pages. Plus is $149 a month, or $74 billed annually, for 10,000 pages.
The result a LlamaParse integration is usually built to produce, in about a minute.
Drag in phone photos, scanned PDFs and emailed invoices, one file or a few hundred at once.
Tip: Start with your worst documents: faded thermal paper, handwritten tips and multi-page vendor bills.
Vendor, date, totals, sales tax and line items are read into consistent columns, with no tier to choose and no schema to write.
Compare values with the images, then download Excel or CSV or send the data to QuickBooks.
Written for US developers, product leads and finance operators deciding between a document parsing API and a finished extraction tool for receipts and invoices.
LlamaParse is built to feed documents into LlamaIndex pipelines. If that is your job, its tiered credits are a fair model. Budget for the tier your hardest pages need, not the cheapest one.
You need client receipts in a spreadsheet today. A parsing API is a component you would have to wire up; a finished tool is the whole job.
Month-end invoice spikes on a credit meter make budgets unpredictable. A flat plan plus a review screen is easier to own than an internal integration.
When someone proposes LlamaParse for invoices, ask for cost per usable record at the tier your documents need, including engineering time.
Every LlamaParse rate on this page comes from the LlamaIndex pricing page and the LlamaParse pricing documentation, in US dollars. Dollar figures per 1,000 pages and per month are our own arithmetic from those published credit rates, and they are labeled that way.
LlamaParse costs $1.25 per 1,000 credits in both North America and Europe. Parsing uses 1 credit a page on Fast, 3 on Cost-effective, 10 on Agentic and 45 on Agentic Plus, which works out to $1.25, $3.75, $12.50 and $56.25 per 1,000 pages. Plans are Free with 10,000 credits a month, Starter at $50 with 40,000 credits and Pro at $500 with 400,000 credits.
Two facts about those plans are worth noticing. First, the included credits are worth exactly the plan price at list rate: 40,000 credits is $50 and 400,000 credits is $500. The subscription buys usage, not a discount; volume discounts on credits are listed for Enterprise only. Second, every plan from Free to Pro lists 100 users, so unlike several document tools, adding a colleague does not force an upgrade.
| Plan | Monthly price | Included credits | Credit value at $1.25 per 1,000 | Pay-as-you-go | Concurrent parse jobs | Projects | Users |
|---|---|---|---|---|---|---|---|
| Free | $0 | 10,000 | $12.50 | Upgrade to Starter | 5 | 1 | 100 |
| Starter | $50 | 40,000 | $50 | Up to $500 a month | 5 | 1 | 100 |
| Pro | $500 | 400,000 | $500 | Up to $5,000 a month | 20 | 5 | 100 |
| Enterprise | Custom | Custom | Volume discount | Custom | 100 | Custom | Custom |
Source: the LlamaIndex pricing page plan cards and comparison table. Enterprise adds SSO, 5x higher rate limits, SaaS or hybrid deployment and a dedicated account manager. Startups can apply to a separate startup program for free credits.
In the current v2 API, LlamaParse uses 1 credit per page on Fast, 3 on Cost-effective, 10 on Agentic and 45 on Agentic Plus. Layout extraction adds 3 credits a page on any tier, and enriched forms output adds 10 credits for each page that contains a form. Spreadsheets bill 1 credit per sheet.
| Parse tier (v2) | Credits per page | Cost per page | Cost per 1,000 pages | Notes from LlamaIndex |
|---|---|---|---|---|
| Fast | 1 | $0.00125 | $1.25 | Spatial text only, no Markdown |
| Cost-effective | 3 | $0.00375 | $3.75 | Recommended starting point for testing |
| Agentic | 10 | $0.0125 | $12.50 | For more complex documents |
| Agentic Plus | 45 | $0.05625 | $56.25 | Highest tier |
| Layout extraction add-on | +3 | +$0.00375 | +$3.75 | Can be added to any tier |
Auto Mode routes each page to a tier and is listed at 3 to 45 credits a page, with LlamaIndex claiming savings of up to 80% against sending everything to a high tier. Re-parsing the same file within 48 hours costs nothing, because parsed files are cached for that window.
The older v1 API priced by mode and model. Its Invoice, Scientific papers, Technical documentation and Forms presets are each listed at 90 credits a page, which is $112.50 per 1,000 pages. Parsing with a large model such as openai-gpt-5 in LVM mode is listed at 150 credits. If an invoice integration was built on v1 presets, checking whether v2 Agentic at 10 credits does the job is the single biggest saving available.
LlamaParse Extract costs 8, 25 or 60 credits a page at its default settings, because the v2 price is the extract tier plus the parse tier underneath it. That is $10, $31.25 or $75 per 1,000 pages. Turbo, which skips the separate parse step, is a flat 35 credits, or $43.75 per 1,000 pages.
| Extract tier | Extract credits | Default parse tier | Parse credits | Total credits per page | Cost per 1,000 pages |
|---|---|---|---|---|---|
| Cost-effective | 5 | Cost-effective | 3 | 8 | $10.00 |
| Agentic | 15 | Agentic | 10 | 25 | $31.25 |
| Turbo | 35, all in | None | 0 | 35 | $43.75 |
| Agentic Plus | 50 | Agentic | 10 | 60 | $75.00 |
Schemas above 250 leaf fields multiply the extract portion: 2x from 251 to 1,000 fields, up to 5x at 3,001 to 3,200. A receipt schema rarely comes close, but a table counts by columns, not rows, so a 40 row, 10 column line item table is 10 fields. Classification costs 1 or 2 credits a page and Split costs 4, which is useful for sorting a mixed upload before paying extraction rates on every page.
Here is the same monthly volume at each default extraction tier, assuming single-page receipts and no retries. The arithmetic is ours from the published credit rates, and it is compared with ReceiptOCR list prices.
| Receipts per month | Extract, Cost-effective (8) | Extract, Agentic (25) | Extract, Agentic Plus (60) | ReceiptOCR |
|---|---|---|---|---|
| 1,000 | $10.00 | $31.25 | $75.00 | $49 Starter, 2,500 pages |
| 2,500 | $25.00 | $78.13 | $187.50 | $49 Starter, or $24 billed annually |
| 10,000 | $100.00 | $312.50 | $750.00 | $149 Plus, or $74 billed annually |
| 100,000 | $1,000.00 | $3,125.00 | $7,500.00 | Custom |
Read that honestly. If your receipts are clean and Cost-effective extraction handles them, LlamaParse is cheaper per page than our Starter plan at 2,500 a month and cheaper than Plus at 10,000. If your receipts need Agentic extraction, the credit bill passes our flat price at around 1,600 receipts a month. Either way the LlamaParse figure is only the model bill. It does not include writing the schema, checking that subtotal plus tax equals the total, building a review screen or exporting to your accounting system.
The Free plan covers a real pilot: 10,000 credits a month is 1,250 receipts at Cost-effective extraction or 400 at Agentic. Starter includes 40,000 credits, which is 5,000 Cost-effective extractions or 1,600 Agentic ones before pay-as-you-go spending begins.
LlamaParse has a Free plan with 10,000 credits a month, worth $12.50 at list rate, 5 concurrent parse jobs, one project and basic community support. To keep parsing past those credits you upgrade to Starter at $50 a month, which unlocks pay-as-you-go usage. It is a free tier for testing, not a free production service.
No. LlamaParse is a hosted, metered service from LlamaIndex. The open source projects LlamaIndex lists are separate: the LlamaIndex framework, Workflows and LiteParse. Enterprise customers can run LlamaParse in a hybrid cloud deployment under a commercial agreement.
No, and several published summaries mix them up. A LlamaParse credit costs $0.00125. A LandingAI credit costs $0.01, which is the "$1 buys 100 credits" line and the Explore plan name you may have seen attached to LlamaParse elsewhere. Neither credit maps to a page one to one, so compare cost per page, never credits. Our Landing AI pricing breakdown converts its credits to dollars per document.
Compared per 1,000 single-page documents, LlamaParse spans the whole market depending on tier. Its Cost-effective extraction is one of the cheaper structured options we have verified, and its Agentic Plus extraction is one of the most expensive.
| Service | What it does | Price per 1,000 pages |
|---|---|---|
| LlamaParse Fast | Spatial text | $1.25 |
| LlamaParse Cost-effective parse | Markdown with tables | $3.75 |
| Mistral OCR 4, standard | Layout-aware text and tables | $4.00 |
| LlamaParse Extract, Cost-effective | Fields to your schema | $10.00 |
| Reducto r-1 Parse | Parsed layout | $10.00 |
| AWS Textract AnalyzeExpense | Prebuilt receipt and invoice fields | $10.00 |
| Reducto Extract | Fields to your schema | $20.00 |
| LlamaParse Extract, Agentic | Fields to your schema | $31.25 |
| LlamaParse Extract, Agentic Plus | Fields to your schema | $75.00 |
| Google Document AI Expense parser | Prebuilt receipt fields, per document | $100.00 |
Rates for the other services are the list prices we verified for Mistral OCR API pricing, Reducto pricing, AWS Textract pricing and Google Document AI pricing. Every meter sits in one table on our OCR API pricing page.
Yes. LlamaIndex lists invoice processing as a use case, Extract returns fields with citations and confidence scores, and Parse can output Markdown, JSON, XLSX and HTML tables. What it does not ship is a finished receipt workflow: you write the schema, pick the tier, validate the numbers and move them into your books. Our invoice OCR API returns prebuilt receipt and invoice fields without that setup.
Buy LlamaParse when you are building retrieval or agent workflows on LlamaIndex, your documents are varied, and you have engineers to tune tiers and schemas. The 100 user allowance and the 48 hour cache make it comfortable for a development team. It also fits the broader trade-off we cover on the LLM OCR page.
Skip it when the goal is receipts and invoices in a spreadsheet for a business, a bookkeeping practice or an AP team. Then the parser is one component and you would be building the rest. Run a batch through the extractor at the top of this page and compare the columns with what a LlamaParse response gives you on the same files.
LlamaParse costs $1.25 per 1,000 credits. Parsing uses 1 to 45 credits a page depending on the tier, or $1.25 to $56.25 per 1,000 pages. Plans are Free with 10,000 credits, Starter at $50 a month with 40,000 credits, Pro at $500 with 400,000 credits, and custom Enterprise.
On the v2 API, Fast uses 1 credit a page, Cost-effective 3, Agentic 10 and Agentic Plus 45. Layout extraction adds 3 credits a page. The older v1 Invoice and Forms presets use 90 credits a page, so check which API version your integration calls.
One LlamaParse credit is worth $0.00125, because 1,000 credits cost $1.25 in both North America and Europe. Included plan credits are valued at the same rate, so the $50 Starter plan buys 40,000 credits and the $500 Pro plan buys 400,000.
Extract bills the extract tier plus the parse tier. At default settings that is 8 credits on Cost-effective, 25 on Agentic and 60 on Agentic Plus, or $10, $31.25 and $75 per 1,000 pages. Turbo skips parsing and costs a flat 35 credits a page.
The Free plan includes 10,000 credits a month, about $12.50 of usage, with 5 concurrent parse jobs and one project. That covers roughly 1,250 receipts at Cost-effective extraction. Continuing past the included credits requires the $50 Starter plan and its pay-as-you-go usage.
No. The Free, Starter and Pro plans each list 100 users on the LlamaIndex pricing page, so seats do not drive the price. Plans differ in credits, pay-as-you-go limits, concurrency, projects and support level.
Not within 48 hours. Parsed files are cached for 48 hours, and re-parsing the same file inside that window costs 0 credits. If a file was already parsed, later extractions pay only the extract tier rate.
At the low tiers, yes. LlamaParse Cost-effective parsing is $3.75 per 1,000 pages against Reducto r-1 Parse at $10, and Cost-effective extraction is $10 against Reducto Extract at $20. At the Agentic and Agentic Plus tiers, LlamaParse extraction costs $31.25 and $75 per 1,000 pages, more than Reducto.
If you want receipt and invoice fields in Excel, CSV or QuickBooks without designing schemas or choosing tiers, a finished extraction tool fits better. ReceiptOCR returns vendor, date, sales tax, total and line items with a review screen, from $49 a month for 2,500 pages.
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
OCR 4 per page and batch rates, with a receipt budget table.
Every published OCR and extraction meter in one per 1,000 page table.
When a language model beats a classic OCR engine, and when it does not.
Enterprise OCR, the Expense parser and the per document trap.
Invoice headers and line items over a REST API.
Parsed receipt fields as JSON, without writing a schema.
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