Mindee
AI document processing API that turns invoices, receipts, IDs and other documents into structured JSON
By Mindee · HQ Paris, France · 4.8/5 Value-Position score (estimate)
Positioning guardrails
Best for
- Accounts payable and finance teams automating invoice, receipt and bank statement capture into JSON
- Product and platform engineers who need a documented REST API with SDKs (Python, Node, Java, PHP) and zero model training
- Fintech and marketplace teams doing instant onboarding, ID, passport and proof-of-residence checks
- Operations teams processing multi-page PDFs (up to 200 pages / 100MB) with split, classify and crop utilities
Ideal size: 10–500 people · Engineering-led scale-up or enterprise with API-first operations
Not for
- Organizations that require fully on-premise or private-cloud deployment with no data leaving their own infrastructure
- Buyers who want a finished end-user application rather than an API to embed in their own product
- Very high-volume processors (30M+ pages/year) where the vendor itself suggests a build-versus-buy review
- Teams wanting a packaged ERP/CRM document suite; connectors are REST, webhook and low-code based
Value metrics scorecard
Time-to-Value
Under a week (API-first, no training)
~7 days to first production value
Total Cost of Ownership
$1,393/yr
Starts at $529 · Credit-based subscription: 1 page = 1 credit (1.5 with a confidence score); monthly or annual billing; overage billed at the plan unit rate plus a 20% surcharge
Implementation Friction
2/5
Engineering + admin effort required
Value-Position score
out of 5 · model estimate
Full cost breakdown
Mandatory implementation fee
None
Seat tiers
Starter includes 1 seat; Pro and Enterprise include unlimited seats
Add-on costs
- Additional credit volumes up to 300,000 credits; overage charged at the plan unit rate plus 20%
- Enterprise custom models, workflow access and platform fees quoted separately
Company & support
Who is behind Mindee, and how your team gets help once it is live.
Company
- Founded
- Not recorded
- Headquarters
- Paris, France
How you get support
- PhoneNot listed
- EmailNot listed
- Live chatPlan not stated
- Support portal / ticketsNot listed
- Community forumAll plans
- Help centre / docsAll plans
- Dedicated account managerEnterprise only
- In person / on-siteNot listed
- Hours
- Not recorded
- Response time
- Not stated
Community support on Starter; live chat from Pro; Enterprise adds a dedicated account manager, premium technical support and custom SLAs. Marketing pages do not publish support hours or a concrete response-time SLA.
“Not listed” means the vendor’s public pages don’t mention that channel, not that it is unavailable. Ask about it during evaluation.
Market position
Where Mindee sits against its closest alternatives. Pick any two of cost, speed, friction and buyer score, and up to 9 companies to compare.
Quadrant view
Typical annual cost × Time-to-value
The lines cross at the median of the solutions shown, so about half sit on each side of each line. A dashed ring marks an outlier pinned to the edge; hover for its value.
Companies on the chart 6 / 10
- Mindee
- OFX
- Exalate
- Hubstaff
- Brex
- greytHR
Add or change companies
Up to 10 companies including Mindee. Listed closest first.
Stack fit signal
Compatibility with standard B2B ecosystems.
No supported MCP path today, so it cannot be driven from an AI client.
AI & MCP readiness
What Mindee ships in AI, and what it asks of your ecosystem.
AI features shipped
Vendor describes pre-trained OCR/extraction models, continuous learning, a RAG knowledge base of past corrections, and per-field confidence scores with bounding boxes. No named foundation-model provider and no BYOK or bundled-model statement appears on the pages reviewed.
Your data & models
- Trains on your data
- Not recorded — ask the vendor
- Runs on
- Not recorded
- AI pricing
- Billed by usage or credits
In your ecosystem
- AI connection
- Not supported
- Model key
- Not recorded
- AI usage audit
- Not recorded
Compliance attestations
* IAPP AIGP certifies individuals, not products. It means named staff hold the credential — not that the platform does.
Bottom line
Mindee is a developer-first AI document processing API. It converts invoices, receipts, passports, bank statements and other documents into structured JSON via a REST API and SDKs, with no model training required. Utilities for splitting, classifying and cropping batch uploads are included. Pricing is credit-based (1 page = 1 credit), starting at $529/year for Starter, with unlimited-seat Pro at $1,393/year. Buyers get confidence scores, polygon coordinates and RAG-based customisation, but must wire the API into their own workflow.
Frequently asked questions
How does Mindee's credit-based pricing work?
Every plan includes the full API; you buy a monthly credit volume. By default 1 processed page = 1 credit, and a page with a confidence score costs 1.5 credits. Utilities such as split, classify and crop are billed at 0.25 credits per page and are waived on pages tied to an extraction template. Credits expire at the end of the subscription period and are not carried over. If you exceed the allowance, Mindee charges overage at the plan unit rate plus a 20% surcharge so production is not blocked.
How long does implementation take and is there a mandatory setup fee?
The vendor positions Mindee as plug-and-play: pre-trained models, zero training time, and SDKs for Python, Java, PHP and more. Mindee's FAQ claims teams can go live in minutes, and there is a 14-day free trial with no credit card. No mandatory implementation fee is published; Enterprise includes 30 minutes of free setup plus 2 hours of solution engineering. Realistically, budget developer time to map fields, validate JSON against your schema and handle exceptions.
What document types does Mindee support out of the box?
Ready-made models cover invoice, receipt, passport, resume, bank statement and driver's licence OCR, among others. The API returns structured JSON with granular confidence scores and bounding boxes, and supports any alphabet or language. A single API call can chain pre-processing utilities (split, classify, crop) with extraction, so multi-page batches, mixed document sets and pages containing several documents can be handled without extra cost.
Can Mindee connect to our ERP, CRM or existing systems?
There is no packaged ERP or CRM connector range. Mindee exposes a REST API that returns JSON, so integration happens through SDKs, low-code HTTP actions, or an automation platform, and inbound webhooks can push results to your endpoint for heavy production use. Mindee also supports asynchronous polling for large files. Buyers needing pre-built SAP, Salesforce or NetSuite apps should plan for custom work or a partner.
Is our document data used to train models, and where is it processed?
The pages reviewed describe localized processing zones and a strict don't-store-my-data retention policy, but they do not explicitly state whether customer data is used to train or improve models, so that must be confirmed with the vendor and the data processing agreement. Processing-zone selection is a Pro-plan feature, which matters to UK and EU buyers with data-residency requirements.
How much does Mindee cost for a mid-market team?
Published list pricing is $529 per year for Starter (1 seat, community support) and $1,393 per year for Pro (unlimited seats, live chat support, data processing localization), both with credit volumes selectable at sign-up and annual billing saving 10%. Enterprise starts at 500,000+ credits per year, is contract-based and is the only tier with a dedicated account manager, custom SLAs and premium technical support. Actual cost tracks page volume, so model your throughput before choosing a tier.
Mindee in our guides
A deliberate starter stack for logistics and field services firms: one pick for each business outcome, the runners-up that beat them in specific cases, and how the pieces fit together.