The short answer
Start with the intake and destination you already use. Prefer a manageable workflow with visible exceptions over extra extraction features your team cannot operate.
A small business usually needs more than text copied out of a PDF. An invoice needs the right supplier and line items; a lease needs dates and rent periods; a customer application needs fields that match the next system. The useful output is a reviewable record with a clear destination.
This shortlist favors practical intake, field definition and delivery over enterprise breadth. Parseur and Docparser are strong starting points for operators, while Nanonets adds workflow controls. Dokyumi, Mindee and Google Document AI become relevant when somebody can own an API integration. Those are different buying decisions, even when every product uses AI.
The order reflects documented fit for a small team, not an accuracy benchmark. We reviewed official capabilities and pricing; we did not upload a shared test set. Choose a pilot around your real document variations before committing to a production workflow.
The shortlist at a glance
Prices use the billing basis shown. Page, task, credit and execution allowances measure different things. On a small screen, scroll the table horizontally.
| Tool | Best fit | Price and billing basis |
|---|---|---|
| 1. Parseur | Email and spreadsheet intake | 20 pages/month free. Micro lists US$49/month for 100 pages, or US$468 billed annually. One credit equals one page. |
| 2. Docparser | Recurring layouts with operator control | Starter: US$39/month for 100 credits and 15 parsers. One credit covers one document up to five pages; longer files use more. |
| 3. Nanonets | Extraction plus business workflow | US$50 starting credits; pricing then lists US$100/month for 100 credits. Usage charges depend on block types and run counts. |
| 4. DokyumiSoxoa-owned product | Custom fields and customer uploads | 25 credits/month free. Starter: US$99/month for 500 credits. One file uses one credit per five pages, rounded up; no automatic overages. |
| 5. Mindee | A developer-owned extraction service | Subscription credits: one per page, or 1.5 with confidence scoring. Default excess usage adds a 20% surcharge; unused credits expire. |
| 6. Google Cloud Document AI | A team already on Google Cloud | Form Parser/Custom Extractor: US$30 per 1,000 pages in the first million/month. Invoice Parser: US$0.10 per document block of 1–10 pages. |
What matters for this decision
- Operator effort
- Can the person receiving documents define fields, inspect exceptions and change the destination without a development project?
- Document variety
- Distinguish stable supplier layouts from changing applications and mixed document classes before choosing rules or a custom schema.
- Usable output
- Check spreadsheet delivery, typed JSON, line items and the treatment of missing values against the next system.
- Predictable capacity
- Price pages, five-page credit blocks and workflow operations separately; a quoted document allowance is not a universal unit.
The ranked picks
Parseur
Best for email and spreadsheet intake
Parseur routes emails and documents through mailboxes with fields defined by AI instructions or layout templates. It exports extracted data to Excel, CSV, JSON and connected applications. That combination fits recurring intake where the same information must reach a spreadsheet or another business system.
Mailboxes, AI fields and templates cover a useful small-team path from incoming email or attachments to spreadsheet data without starting with application code.
The tradeoff: Base plans are single-user; count attachments and PDF pages when sizing the allowance.
Price and billing: 20 pages/month free. Micro lists US$49/month for 100 pages, or US$468 billed annually. One credit equals one page.
Check before choosing: Shared multi-user access requires Scale plans starting at 10,000 credits/month.
Docparser
Best for recurring layouts with operator control
Docparser combines parsing rules, OCR and document-specific filters to extract fields and repeating tables from business documents. It supports direct Excel, CSV, JSON and XML exports alongside an HTTP API and webhooks. Its layout and formatting controls suit repeatable documents that an operator can configure and maintain.
Choose it when consistent forms or supplier documents justify explicit extraction rules, table controls and direct exports that an operator can maintain.
The tradeoff: Check whether your layouts need paid multi-layout support or Professional table features.
Price and billing: Starter: US$39/month for 100 credits and 15 parsers. One credit covers one document up to five pages; longer files use more.
Check before choosing: Multi-layout parsing costs extra on Starter/Professional; Smart Tables begins on Professional.
Nanonets
Best for extraction plus business workflow
Nanonets assembles document workflows from extraction, transformation, lookup, review and export blocks. Review rules can stop a file for correction or approval before downstream delivery. It is useful when capturing text is only one step and the team also needs to validate, route and integrate the result.
The additional value is the sequence around extraction: lookups, transformations, review stages and export can live together instead of becoming separate manual steps.
The tradeoff: Estimate the complete workflow; per-page and per-row runs can change the cost.
Price and billing: US$50 starting credits; pricing then lists US$100/month for 100 credits. Usage charges depend on block types and run counts.
Check before choosing: Extraction runs per page; field and row operations can add runs and cost.
Dokyumi
Best for custom fields and customer uploads · Soxoa-owned
Dokyumi returns JSON shaped by a saved custom schema, with model-reported field confidence and validation details. Intake can use its dashboard, REST API or branded customer upload portals. The application receiving the result keeps control of business checks and approval; a processing status does not approve a transaction.
Its saved schema and branded portals suit a small software or operations team collecting different business documents into its own defined data contract.
The tradeoff: Someone must own downstream integration and approval; this is a Soxoa-owned product.
Price and billing: 25 credits/month free. Starter: US$99/month for 500 credits. One file uses one credit per five pages, rounded up; no automatic overages.
Check before choosing: Self-service files stop at 50 pages; confidence still needs document-specific evaluation.
Mindee
Best for a developer-owned extraction service
Mindee provides extraction APIs built around editable data schemas for catalog or custom document types. Schemas support typed fields, objects and arrays, allowing developers to shape the output for their application. Optional confidence features add review signals, while the surrounding intake, approval and business workflow remains an integration decision.
Editable schemas make it a candidate when a developer can embed extraction inside an existing product rather than introduce a separate operator workflow.
The tradeoff: Account for schema constraints, confidence credits and default excess-usage billing.
Price and billing: Subscription credits: one per page, or 1.5 with confidence scoring. Default excess usage adds a 20% surcharge; unused credits expire.
Check before choosing: Schema nesting is limited; documentation recommends no more than 25 fields for latency.
Google Cloud Document AI
Best for a team already on Google Cloud
Google Document AI separates document processing into purpose-specific processors. Invoice Parser extracts invoice fields and line items; Form Parser returns generic fields and tables; Custom Extractor supports a defined schema. This gives an engineering team model choices, but those choices also change configuration, limitations and the billing unit.
Purpose-specific processors give a technically supported business options for invoices, forms and custom fields without treating all documents as one model.
The tradeoff: Cloud configuration and processor-specific behavior add operating work for a small team.
Price and billing: Form Parser/Custom Extractor: US$30 per 1,000 pages in the first million/month. Invoice Parser: US$0.10 per document block of 1–10 pages.
Check before choosing: Form Parser cannot be up-trained and targets simple tables; processor/version limits differ.
Choose by the constraint you have
- Documents arrive by email; results go to a sheet: ParseurPilot the mailbox route first, including attachment handling and one correction cycle.
- The same layout appears every week: DocparserTest explicit rules before paying for a broader integration platform.
- Your software needs its own JSON shape: Dokyumi or MindeeCompare schema expressiveness, limits and review behavior on the same permitted files.
Run a pilot that includes the awkward files
- Choose representative documents from every recurring layout, including scans, multi-page files and documents with missing fields.
- Write expected values before running the pilot. Score required fields and complete records separately; count incorrect nonempty values as errors.
- Have the intended operator correct exceptions and deliver the result to the real destination. Record time and unresolved cases.
- Calculate cost using actual page lengths and workflow operations, then set a review rule before connecting live business actions.
Keep the trial result and the original input together. A vendor feature list establishes what to investigate; it does not establish your accuracy, savings or operating cost.
Questions before you commit
Is document extraction the same as OCR?
OCR recognizes text. Extraction organizes selected information into fields, tables or another data contract. You may still need to validate totals, identities and dates after extraction; readable text alone does not establish that a record is correct.
Should a small business choose templates or AI?
Stable layouts can justify explicit templates and rules. Variable layouts may favor AI fields or a custom schema. Try both against the documents that cause corrections; the product label does not establish which route needs less work for your team.
Can we compare advertised prices per document?
Only after defining the document. A ten-page file can consume ten page credits, two five-page blocks, several workflow runs or a specialized parser charge. Include review, integration and unused subscription capacity in the business decision.
Sources and research limits
We checked the official pages linked below on October 27, 2026. Recommendations combine those documented capabilities with the specific workflow in this guide. Features, rates and availability can change. Confirm the current terms and run the evaluation above before choosing.
Published prices are in US dollars unless explicitly stated. An annual-billing monthly equivalent requires the vendor's annual commitment. A quote-based plan is not treated as free. Missing or ambiguous prices are identified instead of estimated.
- Parseur. Features and pricing · Credit accounting. Checked October 27, 2026.
- Docparser. Extraction features · Plans and add-ons · Cloud-only deployment. Checked October 27, 2026.
- Nanonets. Plans and block rates · How runs are billed · Review stages. Checked October 27, 2026.
- Dokyumi. Workflow and output contract · Plans, credits and limits. Checked October 27, 2026.
- Mindee. Data schemas and constraints · Extraction models · Credit and overage terms. Checked October 27, 2026.
- Google Cloud Document AI. Processors and supported fields · Form Parser limitations · Processor-specific pricing. Checked October 27, 2026.