Reducto
Reducto is a San Francisco-based AI company, founded in 2023, that provides an agentic document platform converting PDFs, images, and spreadsheets into structured, LLM-ready data via Parse, Extract, Split, Classify, and Edit APIs, serving AI teams in finance, healthcare, legal, and insurance.
- Company typePrivate
- Founded2023
- HeadquartersSan Francisco, United States
- Headcount51–100
- GTM typeB2B
- OfferingSoftware
What Reducto does
Reducto is a San Francisco-based, venture-backed software company founded in 2023 that operates an agentic document platform for AI teams. Its product converts unstructured documents — PDFs, scanned images, spreadsheets, presentations, and Office files — into LLM-ready structured data, with five core API endpoints (Parse, Extract, Split, Classify, Edit) layered over a no-code Studio visual workbench. The company's underlying technology combines layout-aware computer vision (OCR) with vision-language models in a multi-pass pipeline in which an agentic model reviews and corrects outputs, producing citation-grounded JSON with bounding boxes and confidence scores; flagship features include Deep Extract (an agent-harness mode reaching 99–100% field accuracy on documents up to 2,500 pages) and Deep Split (an agent-harness document segmentation mode).
The platform is sold via a freemium, usage-based credit model (15K free credits, then $0.015 per credit) on the Standard tier, with custom-priced Growth subscriptions and Enterprise contracts that include SOC 2 Type II, HIPAA/BAA, EU/AU data residency, VPC/on-prem deployment, and dedicated support. Distribution combines product-led self-serve (Studio, docs, Playground), direct enterprise sales supported by Forward Deployed Engineers, an AWS Marketplace listing (since February 2026), and documented integrations with Databricks and Elasticsearch. Named customers span finance (Top 5 Global Hedge Fund, Benchmark, Legora, LEA), legal (Harvey, August, Supio), healthcare (Anterior, Medallion), insurance (Elysian, Newfront), AI infrastructure (Scale AI, Vanta, Stack AI, Mercor, Gumloop), and large enterprise (Toast, JLL, a Fortune 10).
Reducto has raised $108M in primary capital across three rounds — an $8.4M seed led by First Round (October 2024), a $24.5M Series A led by Benchmark (April 2025), and a $75M Series B led by Andreessen Horowitz at a reported $600M valuation (October 2025) — with the capital earmarked for product development, enterprise expansion, and a May 2026 acquisition of edtech startup Opennote. The company reports 3B+ pages processed cumulatively across its platform and has been recognized on Forbes' 2025 Next Billion-Dollar Startups list, Redpoint Ventures' 2026 InfraRed 100, and Forbes' 2026 30 Under 30 (co-founders Adit Abraham and Raunak Chowdhuri).
Reducto firmographics
Firmographics- Name
- Reducto
- Legal name
- Reducto, Inc.
- Website
- https://reducto.ai
- Company type
- Private
- Founded year
- 2023
- Operating status
- Operating
- Headcount range
- 51–100 employees
- Short description
- Reducto is a San Francisco-based AI company, founded in 2023, that provides an agentic document platform converting PDFs, images, and spreadsheets into structured, LLM-ready data via Parse, Extract, Split, Classify, and Edit APIs, serving AI teams in finance, healthcare, legal, and insurance.
- Ownership category
- akta.pro rank
Reducto industry classification
Industry- Product category
- AI Document Intelligence
- NAICS
- Custom Computer Programming Services (541511), Software Publishers (5132), Document Preparation Services (561410)
- SIC
- Services-Prepackaged Software (7372), Services-Computer Programming, Data Processing, Etc. (7370)
- akta.pro primary industry
- Document Capture, Scanning, OCR/ICR & Intelligent Document Processing (IDP) (HDAEAGAD)
- akta.pro secondary industry
- Process & Case Management Low-Code Platforms (HDAEAKAE)
Keywords
Where Reducto is headquartered
LocationHeadquarters
- HQ city
- San Francisco
- HQ country
- United States
- HQ region
- North America
Offices2 records
Markets served
Reducto business model
Business model- GTM type
- B2B
- Offering type
- Software
- Cost components
- Technology or R&D, Personnel, Infrastructure, Marketing or Sales, Operations
Revenue model
- Usage-based API consumption (credits): Pay-as-you-go credit consumption on Standard tier ($0.015/credit after the first 15K free credits). Every parse/extract/split/edit call consumes credits based on operation type, configurations, and pages processed.
- Freemium tier (Standard free credits): Free Standard tier with up to 15K credits to drive developer adoption and self-serve conversion into paid usage.
- Growth tier subscription: Custom-priced subscription tier for scaling teams, including volume discounts, priority support, BAA, zero data retention, and EU/AU data residency.
- Enterprise contracts: Custom MSA/SLA contracts with VPC/on-prem deployments, custom rate limits, RBAC, SSO/SAML, dedicated on-call, and custom processing pipelines for regulated/enterprise customers.
- AWS Marketplace channel: Available on AWS Marketplace since February 2026, enabling enterprises to purchase using committed AWS spend (transaction-fee/marketplace channel).
- Professional services / Forward Deployed Engineering: Forward Deployed Engineer roles suggest hands-on deployment services bundled with enterprise contracts to support complex customer integrations.
Pricing tiers
| Model | Billing | Price |
|---|---|---|
| Freemium | Pay-as-you-go | Standard — free up to 15K credits, then $0.015/credit pay-as-you-go |
| Subscription | Annual | Growth — custom pricing for scaling teams with volume discounts |
| Subscription | Multi-year contract | Enterprise — custom pricing with full control and security features |
Go-to-market motion5 records
Distribution channels5 records
Marketing channels10 records
Reducto product offering
Product offeringCore offering
Reducto offers an agentic document platform composed of five core REST API endpoints—Parse, Extract, Split, Classify, and Edit—plus a no-code visual Studio workbench. The platform converts unstructured documents (PDFs, images, spreadsheets, presentations, DOCX, handwriting, charts) into structured, LLM-ready JSON with citation-grounded bounding-box output. Customers consume the platform via self-serve Studio, direct enterprise contracts, or AWS Marketplace, with deployment options ranging from multi-tenant cloud to VPC and on-prem.
Product overview
Reducto's offering is a single unified agentic document platform composed of five core API endpoints — Parse, Extract (with the Deep Extract agent-harness mode), Split (with Deep Split), Classify, and Edit — plus a no-code visual Studio workbench that lets teams design, test, version, and deploy document pipelines by Pipeline ID. The endpoints fit together as a composable pipeline: Classify routes incoming files by type, Split segments multi-document files into page ranges, Parse produces the structured LLM-ready JSON backbone, Extract pulls schema-typed fields (with Deep Extract as an agent-verification mode for long, complex documents), and Edit writes data back into PDF forms and DOCX files. Studio sits on top of the same engine as the API and is included with every plan, while Smart Schema is a Studio add-on that auto-generates extraction schemas. The platform is targeted at AI teams in finance, healthcare, insurance, legal, government, and logistics, with deep integrations into RAG and search pipelines (Elasticsearch, Databricks), enterprise security (SOC 2 Type II, HIPAA), and procurement channels (AWS Marketplace).
Differentiator
Problem solved
Functional benefit
Products and services
- Parse API REST endpoint that turns any document (PDF, image, spreadsheet, Office file) into structured, LLM-ready JSON with OCR, layout detection, table reconstruction, figure summarization, and semantic chunking in a single call. Each block returns with type, page position, bounding box, and confidence score.
- Extract API REST endpoint that returns schema-typed JSON fields from any document using a user-defined schema, with optional citations on every value (page, bounding box, source text, confidence). Runs Parse internally and uses an LLM to locate and pull values.
- Split API REST endpoint that classifies every page of a document against user-defined sections described in natural language and returns the page numbers each section occupies, with optional partition_key for repeating sections and high/low confidence per section.
- Classify API REST endpoint that labels a document against a user-defined list of categories with per-criterion and per-category confidence scores; synchronous and lightweight so it can run at the top of a pipeline before heavier Parse or Extract calls.
- Edit API REST endpoint that writes data into PDF forms (with or without AcroForm widgets) and modifies DOCX files using natural-language instructions or a reusable form_schema. Fills text, checkbox, and dropdown fields; highlights changes in DOCX; supports overflow pages for long values.
- Studio No-code visual environment at studio.reducto.ai for designing, testing, versioning, and deploying document pipelines by Pipeline ID. Runs on the same engine as the API, with side-by-side citations, bounding boxes, and rollback/audit logging on Enterprise.
Quantifiable outcome
- 3B+ pages processed across the platform
- +8 more outcomes
Companies that use Reducto
Customer profileNamed customers24 records
Segments7 records
Ideal customer profiles7 records
Reducto technology and API
TechnologyTechnology focussed Yes
API detail
- Has API
- Yes
- API docs
- API detail
Core technology
AI maturity
App detail
Integration2 records
AI capability10 records
Feature11 records
Reducto partnerships and signals
Strategic signalPartnerships
Four partnerships are on record, tiered core.
- OpennotecoreReducto acquired Canadian-founded edtech startup Opennote (Y Combinator Summer 2025 batch) in May 2026. Opennote's entire team joined Reducto to enhance document parsing and workflow capabilities; Opennote's consumer AI software offerings are being sunset.
- AWS MarketplacecoreReducto is available on AWS Marketplace, enabling enterprises to purchase the platform using committed AWS spend. Announced February 6, 2026.
- DatabrickscoreCo-published blog content on streamlining document processing with Reducto and Databricks; integration enables joint customers to unlock unstructured data inside Databricks.
- ElasticsearchcorePublished tutorial on using Reducto parsing with Elasticsearch for semantic search; integration enables RAG and search workflows over Reducto-parsed documents.
Scale indicators18 records
Recent moves5 records
Expansion highlights7 records
Reducto competitors and assessment
Company assessmentEmerging players
- Rossum: Rossum is an AI-first invoice and document processing platform with a strong AP/invoice-automation focus. It overlaps with Reducto on enterprise document extraction and classification workflows, though it leans more toward transactional finance than broad LLM-ingestion.
- Mistral OCR: Mistral's OCR endpoint is a frontier-model document understanding API published directly by a foundation-model vendor. It represents the emerging threat of base-model providers shipping first-party document capabilities that compete with Reducto's parse/extract offerings.
- Mathpix: Mathpix converts PDFs, images, and handwriting (especially scientific/STEM) into structured data via OCR and LLMs. It is an emerging adjacent player focused on higher-fidelity extraction of equations, tables, and text, with partial overlap to Reducto's academic/financial document use cases.
Broad incumbents
- Adobe Document Services: Adobe's Document Services (PDF Extract, PDF Services API) provide structured extraction and manipulation of PDFs as part of Adobe's broader document cloud. It is a broad incumbent alternative for organizations already using Adobe's PDF stack who might otherwise use Reducto.
- AWS Textract: AWS Textract is the hyperscaler incumbent for OCR and document extraction inside the AWS ecosystem. It competes on forms, tables, and handwriting at the platform layer rather than as a specialized AI-infrastructure product, and is often an alternative to Reducto inside AWS-native shops.
- Azure AI Document Intelligence: Microsoft's Azure AI Document Intelligence (formerly Form Recognizer) offers prebuilt and custom document models integrated with the Azure AI stack. It is a broad incumbent alternative for Microsoft-centric enterprises that might otherwise evaluate Reducto for Forms, invoices, and custom extraction.
- Google Document AI: Google Cloud's Document AI provides OCR, form parsing, and custom document extractors as part of a broader AI platform. It is a broad incumbent alternative to Reducto for enterprises standardized on GCP, with overlapping accuracy/throughput claims on structured and unstructured documents.
Direct peers
- Nanonets: Nanonets provides an AI-based document extraction and workflow automation platform with OCR, classification, and API endpoints for enterprise teams. It is a direct competitor to Reducto's Extract/Classify endpoints, particularly in finance, insurance, and operations use cases.
- Unstructured: Unstructured offers a document parsing/ingestion API aimed at LLM and RAG pipelines, the closest direct competitor to Reducto's Parse/Extract endpoints. Both target AI engineering teams converting messy enterprise documents into LLM-ready structured output.
- LlamaParse (LlamaIndex): LlamaParse is LlamaIndex's document parsing service for complex PDFs, tables, and figures used in RAG pipelines. It overlaps directly with Reducto's Parse API and is similarly positioned for AI engineers building retrieval and agent applications.
Market position
Strengths5 records
Weaknesses5 records
Competitive moat5 records
Key risks6 records
Key highlights7 records
Customer concentration
Reducto social profiles
Digital presenceReducto compliance and trust
Trust signalCompliance3 records
Reducto financial estimates
Financial estimateRevenue estimate
Valuation estimate
Reducto leadership team
Management profileNumber of profiles
Profiles2 records
Reducto subsidiaries and ownership
Company hierarchySubsidiaries1 record
Reducto funding detail
Funding detailFunding overview
Funding rounds4 records
Investors7 records
Funding detail is available on the Subscription and Enterprise plan.Contact sales →
Reducto M&A and investment
M&A and investmentM&A1 record
Investments
M&A and investment is available on the Subscription and Enterprise plan.Contact sales →
Frequently asked questions about Reducto
What does Reducto do?
Reducto offers an agentic document platform composed of five core REST API endpoints—Parse, Extract, Split, Classify, and Edit—plus a no-code visual Studio workbench. The platform converts unstructured documents (PDFs, images, spreadsheets, presentations, DOCX, handwriting, charts) into structured, LLM-ready JSON with citation-grounded bounding-box output. Customers consume the platform via self-serve Studio, direct enterprise contracts, or AWS Marketplace, with deployment options ranging from multi-tenant cloud to VPC and on-prem.
Is Reducto a public or private company?
Reducto is a private company. It is classified as venture growth investor backed and is currently operating.
When was Reducto founded?
Reducto was founded in 2023. It employs 51 to 100 people.
Where is Reducto based?
Reducto is headquartered in San Francisco, United States, in the North America region.
How does Reducto make money?
Six revenue lines are on record. Usage-based API consumption (credits) is the primary driver. The others are freemium tier (Standard free credits), growth tier subscription, enterprise contracts, AWS Marketplace channel and professional services / Forward Deployed Engineering.
Who are Reducto's main competitors?
Emerging players on record are Rossum, Mistral OCR and Mathpix. Broad incumbents are Adobe Document Services, AWS Textract, Azure AI Document Intelligence and Google Document AI. Direct peers are Nanonets, Unstructured and LlamaParse (LlamaIndex).
Does Reducto have an API?
Yes. Reducto's public API exposes five core REST endpoints: /parse (turns documents into structured, LLM-ready JSON with OCR, layout detection, table reconstruction, figure summarization, and semantic chunking), /extract (returns schema-typed JSON fields with optional citations, including a Deep Extract agent-harness mode), /split (classifies pages and returns section page ranges with optional partition keys), /classify (routes documents by user-defined categories with per-criterion confidence), and /edit (fills PDF forms and modifies DOCX files via natural-language instructions). The API supports both synchronous low-latency calls and asynchronous batch jobs with custom webhooks. Files up to 5GB can be uploaded via presigned URL, and parsed results can be reused across subsequent calls via a jobid:// reference to skip re-processing. Authentication is via API keys issued through the Reducto Studio account system (studio.reducto.ai). Public documentation is hosted at docs.reducto.ai covering all five endpoints and Studio. Developer documentation is at docs.reducto.ai.
What industry is Reducto in?
Reducto's product category is AI Document Intelligence. Its primary akta.pro industry code is HDAEAGAD, Document Capture, Scanning, OCR/ICR & Intelligent Document Processing (IDP), with a secondary code of HDAEAKAE, Process & Case Management Low-Code Platforms. Its NAICS code is 541511 and its SIC code is 7372.