# Acuity Extract

*/Startups/Acuity_Extract*

## Startup Overview

This document processing engine parses complex visual documents directly into structured, queryable JSON schemas. It handles dense layouts, irregular tables, and nested information without requiring pre-defined templates or coordinate mapping.

Data engineers and operations teams rely on structured data trapped inside inconsistent, high-density files like shipping manifests and financial statements. When document formats shift, legacy extraction tools break. This forces organizations to fall back on slow manual BPO data entry or dedicate continuous engineering time to rebuild brittle bounding-box rules.

Instead of the rigid setup and probabilistic text dumps provided by tools like AWS Textract or Abbyy FlexiCapture, this approach is schema-agnostic on document setup and mathematically deterministic on output. Users define the required data structure, and the system maps unstructured visual inputs to those exact JSON keys, delivering production-ready data ready for immediate database ingestion.

## Startup Founding Hypothesis

**Approach**: that parses complex visual documents into queryable JSON schemas
**Competitors**:
- [AWS Textract](/Competitors/AWS_Textract)
- [Abbyy FlexiCapture](/Competitors/Abbyy_FlexiCapture)
- [Manual BPO Data Entry](/Competitors/Manual_BPO_Data_Entry)
**Differentiator2x2**: schema-agnostic on document setup and mathematically deterministic on output

## Startup Solution Coordinate

**Solution**: [Acuity Extract Engine](/Software/Acuity_Extract_Engine)

## Startup Position2x2

```mermaid
quadrantChart
title Document Parsing Defensibility
x-axis "Template-Dependent" --> "Schema-Agnostic"
y-axis "Probabilistic" --> "Mathematically Deterministic"
quadrant-1 "Defensible Automation"
quadrant-2 "Rigid Legacy"
quadrant-3 "Stagnant Ground"
quadrant-4 "Unreliable Scale"
Manual BPO Data Entry: [0.85, 0.15]
Abbyy FlexiCapture: [0.15, 0.85]
AWS Textract: [0.60, 0.35]
Acuity Extract: [0.90, 0.90]
```

## Startup Offer

**Proof**:
- Targeting 99.9% structural accuracy on multi-page logistics documents without requiring pre-built templates.
- Aiming to eliminate the custom parsing scripts engineering teams maintain to clean up raw OCR text dumps.
- Intended to map highly-variable supplier invoices to strict internal database schemas in under 3 seconds per page.
**Tiers**:
- Name: Developer Build · Price: ~$0.04–$0.06 per document · Inclusions: Up to 5,000 visual document parses per month, standard REST API access, and basic JSON schema validation capabilities.
- Name: Production Scale · Price: ~$0.01–$0.03 per document · Inclusions: Volume tier for 5,000 to 100,000 document parses per month, enabling complex nested schema mapping and direct webhook integrations.
- Name: Enterprise Node · Price: ~$2,500–$6,000/mo minimum commitment · Inclusions: Dedicated processing throughput for over 100,000 documents, mathematically deterministic output SLAs, and priority engineering support.
**Guarantee**: If a document parse returns invalid JSON or violates your strictly defined schema structure, that API call is flagged and not billed.
**Business Function**: ProvideService
**Objection Handlers**:
- Concern: Generative models hallucinate financial numbers. Rebuttal: Acuity Extract enforces mathematically deterministic schema bounds, returning a strict null or error flag instead of guessing ambiguous values.
- Concern: Our vendors change document formats constantly. Rebuttal: The API is schema-agnostic on setup; it uses spatial reasoning to map the target JSON rather than relying on brittle, hard-coded bounding boxes.
- Concern: The per-page cost is higher than basic OCR APIs. Rebuttal: Raw OCR requires expensive downstream engineering to structure the text; this delivers query-ready JSON immediately, eliminating the hidden secondary processing costs.
**Pricing Architecture**: UsageMeter
**Agent Checkout Support**:
- agentic-commerce-protocol

## Startup Brand

**Voice**: Technical and precise, prioritizing mathematical accuracy over marketing fluff.
**Tagline**: Exact JSON schemas from complex visual documents.
**Icon Concept**: Invoice
**Palette Intent**: electric-signal
**Visual Identity**: The visual identity contrasts sharp terminal green against matte charcoal, using monospaced typography to emphasize structured document extraction.
**Archetype Reference**: the-sage

## Startup Buyer Chain

**Chain**: Acuity Extract → Data Engineers → Enterprise Operations Teams
**Gtm Motion**: Acquires technical users via a self-serve API sandbox that lets developers test deterministic extraction on their own complex visual documents. Expands revenue through usage-based, per-page volume tiers as engineering teams route additional document schemas through the API in production workflows.
**Agent Channel**: Designed to list in the LangChain tool registry and OpenAI schema directories as a document parsing utility, enabling autonomous AI agents to discover and invoke the extraction service when encountering unstructured visual files.
**Primary Channel**: Technical SEO and developer documentation targeting high-intent search queries like 'deterministic PDF to JSON API' and 'AWS Textract alternative for complex tables', capturing developers actively searching for parsing solutions.

## Startup Customer Journey

```mermaid
flowchart LR; A[Developer Documentation] --> B[API Sandbox]; B --> C[JSON Schema Output]; C --> D[Production Data Pipeline]; D --> E[Volume Pricing Tier]; E --> F[Tool Registry Listing];
```

## Startup Proof Points

_Illustrative — target and order-of-magnitude estimate figures, not an achieved track record (this Thing is concept-stage)._

**Pilot Goals**:
- 30-day parallel processing pilot: Run 10,000 legacy OCR documents through Acuity Extract alongside the existing system to prove 100% adherence to the client's strict database schema without a single hallucinated financial value.
- 14-day template elimination sprint: Replace a client's hard-coded bounding box system for 10 highly-variable vendor invoice formats using a single schema definition, aiming to prove a 90% reduction in new-vendor onboarding time.
**Target Metrics**:
- Target: 99.9% structural accuracy rate on multi-page logistics documents without pre-built templates
- Aim: 0% unflagged JSON schema violations on complex nested document structures
- Target: <3 seconds processing latency per highly-variable supplier invoice page
- Aim: 100% elimination of hard-coded OCR bounding box maintenance tickets for engineering teams
**Target Case Studies**:
- Target: A mid-market freight forwarding company. Transformation: Replace 10+ custom OCR parsing scripts with a single API call, eliminating the engineering hours previously spent maintaining templates for constantly changing customs document formats.
- Target: An enterprise Accounts Payable automation provider. Transformation: Ingest highly-variable supplier invoices and map the financial data directly to strict internal database schemas in under 3 seconds per page, without requiring manual template creation for new vendors.
- Target: A regional healthcare claims processor. Transformation: Extract complex tabular data from multi-page medical bills into query-ready JSON, relying on spatial reasoning to completely bypass brittle bounding-box OCR setups.
**Testimonial Targets**:
- VP of Engineering: Relief that their development team no longer spends weekly sprints maintaining brittle Regex rules and custom parsing scripts to clean up raw OCR text dumps.
- Lead Backend Developer: Excitement that the API strictly adheres to nested schema mapping and actually returns query-ready JSON exactly as defined without structural errors.
- Head of Finance Operations: Confidence that the spatial reasoning engine flags ambiguous financial numbers or strictly enforces deterministic bounds instead of hallucinating incorrect invoice totals.

## Startup Top Risks

**Risks**:
- Severity: existential · Description: AWS Textract updates its core models to support dynamic, schema-agnostic extraction, rendering the primary competitive differentiator obsolete. · Mitigation Status: unmitigated
- Severity: high · Description: Complex document layouts containing handwritten or heavily degraded text break the mathematically deterministic output guarantees. · Mitigation Status: in-progress
- Severity: moderate · Description: Customer switching costs from entrenched manual BPO contracts and existing Abbyy FlexiCapture templates slow enterprise sales cycles past twelve months. · Mitigation Status: in-progress
- Severity: moderate · Description: Legacy enterprise clients demand fully air-gapped deployments for data privacy compliance, stretching engineering resources away from the core API. · Mitigation Status: unmitigated

## Startup Competitors

- [AWS Textract](/Competitors/AWS_Textract) — Cloud Incumbent
- [Abbyy FlexiCapture](/Competitors/Abbyy_FlexiCapture) — Legacy OCR
- [Manual BPO Data Entry](/Competitors/Manual_BPO_Data_Entry) — Status Quo
- [Google Document AI](/Competitors/Google_Document_AI) — Cloud Incumbent
- [Sensible Platform](/Competitors/Sensible_Platform) — IDP Startup

## Startup Solution Stack

- [Visual Extraction Service](/Services/Visual_Extraction_Service) — Service-as-Software
- [Deterministic Parsing Agent](/Agents/Deterministic_Parsing_Agent) — Agent
- [Schema Generation Worker](/Agents/Schema_Generation_Worker) — Agent
- [JSON Formatting API](/Software/JSON_Formatting_API) — Software
- [Layout Mapping Engine](/Software/Layout_Mapping_Engine) — Software

## Startup Story Brand

**Hero**:
- **Need**: to be the systems architect who builds resilient integrations, not a maintainer of brittle OCR scripts
- **Want**: to extract structured JSON data from variable supplier documents
- **Identity**: the lead engineer at a logistics or fintech firm
**Plan**:
- Step: Define Schema · Detail: Provide the JSON structure your database requires without building custom document templates.
- Step: Audit Output · Detail: Review the deterministic extraction results to ensure every nested field aligns with your strict schema logic.
- Step: Scale Volume · Detail: Push thousands of multi-page documents through the REST API with per-page usage pricing.
**Guide**:
- **Empathy**: You shouldn't still be debugging bounding box failures. ABBYY FlexiCapture wasn't built to handle the infinite variability of modern supplier invoices.
**Problem**:
- **Villain**: unstructured OCR blobs
- **External**: Engineers spend hundreds of hours writing regex and cleanup scripts to fix broken AWS Textract outputs and raw text dumps.
- **Internal**: You feel like a glorified data cleaner instead of a developer building high-value features.
- **Philosophical**: Visual intelligence was built for structural understanding, not just dumping characters onto a page.
**Success**: Engineering teams eliminate document-cleaning debt and deliver query-ready data to internal databases in seconds.
**One Liner**: Every day, engineering leads waste hours cleaning raw OCR text. Acuity_Extract parses complex visual documents into queryable JSON schemas so you get structured data without maintenance scripts.
**Positioning**:
- **So That**: you get valid JSON instead of raw text blobs
- **Unlike**: AWS Textract and manual BPO entry
- **For Whom**: logistics and fintech engineering leads
- **Category**: Automated Document Extraction API
**Call To Action**:
- **Direct**: Parse a document
- **Transitional**: Download sample JSON schemas
**Failure Stakes**:
- Hallucinated financial data
- Constant script maintenance
- Delayed downstream processing
**Transformation**:
- **To**: free to architect autonomous data pipelines, no longer fixing broken OCR strings
- **From**: the developer patching brittle AWS Textract regex
**Controlling Idea**: Document extraction must be mathematically deterministic to be useful for production engineering.

## Startup Token Hero

**Genre**: founding-hypothesis
**Rendered**: Every day, engineering leads waste hours cleaning raw OCR text. Acuity_Extract parses complex visual documents into queryable JSON schemas so you get structured data without maintenance scripts.
**Mechanism**: spine-derived-v1
**Template Id**: spine-founding-hypothesis
**Vocab Fingerprint**: 10bea225867005a2

## Startup Token Positioning

**Genre**: moore-positioning
**Rendered**: Automated Document Extraction API for logistics and fintech engineering leads. Unlike AWS Textract and manual BPO entry — you get valid JSON instead of raw text blobs.
**Mechanism**: spine-derived-v1
**Template Id**: spine-moore-positioning
**Vocab Fingerprint**: 49c7c71d87873522

## Startup Token Pitch Deck

**Genre**: pitch-deck
**Rendered**: Problem: Engineers spend hundreds of hours writing regex and cleanup scripts to fix broken AWS Textract outputs and raw text dumps.
Solution: Every day, engineering leads waste hours cleaning raw OCR text. Acuity_Extract parses complex visual documents into queryable JSON schemas so you get structured data without maintenance scripts.
Customer: logistics and fintech engineering leads
Unlike: AWS Textract and manual BPO entry
**Mechanism**: spine-derived-v1
**Template Id**: spine-pitch-deck
**Vocab Fingerprint**: 71056b1c54e0c11d

## Startup Token M E D D P I C C

**Pain**: Engineers spend hundreds of hours writing regex and cleanup scripts to fix broken AWS Textract outputs and raw text dumps.
**Metrics**: Target: Engineering teams eliminate document-cleaning debt and deliver query-ready data to internal databases in seconds.
**Rendered**: Pain: Engineers spend hundreds of hours writing regex and cleanup scripts to fix broken AWS Textract outputs and raw text dumps.
Economic buyer: Data Engineers
Metrics: Target: Engineering teams eliminate document-cleaning debt and deliver query-ready data to internal databases in seconds.
Competition: AWS Textract and manual BPO entry
**Mechanism**: spine-derived-v1
**Competition**: AWS Textract and manual BPO entry
**Economic Buyer**: Data Engineers
**Vocab Fingerprint**: 969f515a49ab2ef4

## Startup Token Cold Email

**Genre**: cold-email
**Rendered**: Subject: Automated Document Extraction API for logistics and fintech engineering leads

logistics and fintech engineering leads — Engineers spend hundreds of hours writing regex and cleanup scripts to fix broken AWS Textract outputs and raw text dumps. Every day, engineering leads waste hours cleaning raw OCR text. Acuity_Extract parses complex visual documents into queryable JSON schemas so you get structured data without maintenance scripts.
**Mechanism**: spine-derived-v1
**Template Id**: spine-cold-email
**Vocab Fingerprint**: 38ab7d84ebab1b5e

## Startup Token Agent Spec

**Genre**: ai-agent-spec
**Rendered**: Automated Document Extraction API. Every day, engineering leads waste hours cleaning raw OCR text. Acuity_Extract parses complex visual documents into queryable JSON schemas so you get structured data without maintenance scripts. Serves logistics and fintech engineering leads.
**Mechanism**: spine-derived-v1
**Template Id**: spine-ai-agent-spec
**Vocab Fingerprint**: bd7f9a69b2c17d7b

## Neighborhood

### Candidate solutions

- [Manual Referral Transcription](/Problems/Manual_Referral_Transcription) — candidate solution for · Problems

### What it offers

- [Acuity Extract Engine](/Software/Acuity_Extract_Engine) — offers · Software

### Composed of

- [Deterministic Parsing Agent](/Agents/Deterministic_Parsing_Agent) — composes · Agents
- [Visual Extraction Service](/Services/Visual_Extraction_Service) — composes · Services
- [Schema Generation Worker](/Agents/Schema_Generation_Worker) — composes · Agents
- [JSON Formatting API](/Software/JSON_Formatting_API) — composes · Software
- [Layout Mapping Engine](/Software/Layout_Mapping_Engine) — composes · Software

### Embodies

- [Software](/Theses/Software) — embodies · Theses

### Competitors

- [AWS Textract](/Competitors/AWS_Textract) — competes with · Competitors
- [Abbyy FlexiCapture](/Competitors/Abbyy_FlexiCapture) — competes with · Competitors
- [Manual BPO Data Entry](/Competitors/Manual_BPO_Data_Entry) — competes with · Competitors
- [Google Document AI](/Competitors/Google_Document_AI) — competes with · Competitors
- [Sensible Platform](/Competitors/Sensible_Platform) — competes with · Competitors

### Similar Startups

- [Visoph](/Startups/Visoph) — similar · Startups
- [Ocviv](/Startups/Ocviv) — similar · Startups
- [Parseaxis](/Startups/Parseaxis) — similar · Startups
- [Nostruct](/Startups/Nostruct) — similar · Startups
- [Paperinsight](/Startups/Paperinsight) — similar · Startups
- [Napot](/Startups/Napot) — similar · Startups
- [Tablelayer](/Startups/Tablelayer) — similar · Startups
- [Amberparsing](/Startups/Amberparsing) — similar · Startups
- [Documentsense](/Startups/Documentsense) — similar · Startups
- [Eonform](/Startups/Eonform) — similar · Startups
- [Exceaver](/Startups/Exceaver) — similar · Startups
- [Contextual Clerk](/Startups/Contextual_Clerk) — similar · Startups
- [Doquint](/Startups/Doquint) — similar · Startups
- [Docapacity](/Startups/Docapacity) — similar · Startups
- [Structity](/Startups/Structity) — similar · Startups
- [Problata](/Startups/Problata) — similar · Startups
- [Strucvert](/Startups/Strucvert) — similar · Startups
- [Struclum](/Startups/Struclum) — similar · Startups
- [Mentica](/Startups/Mentica) — similar · Startups
- [Tractablenon](/Startups/Tractablenon) — similar · Startups