Unsiloed AI Blog · May 2026
Company blog post announcing Unsiloed Parser v3.1 ranked #1 on olmOCR-Bench, with reproducible benchmark code and detailed evaluation against major OCR/document-AI vendors.
Company research
Public company, workplace, funding, and market signals
Updated Jul 29, 2026
Overview
Unsiloed AI is a San Francisco-based document-AI infrastructure startup that turns PDFs, images, spreadsheets, scans, Word/PPT files, and other multimodal documents into structured JSON and Markdown for LLMs and AI agents. Its primary product is a document-processing API/layer for parsing, extraction, splitting, and classification, aimed at accuracy-sensitive enterprise workflows in finance, legal, and healthcare. Public sources disagree on whether the company was founded in 2024 or 2025, but the balance of evidence supports 2024; the company is very small and founder-led.
Primary product
Document-processing API/layer that converts multimodal documents into LLM-ready JSON and Markdown (parse, extract, split, classify).
Founded
2024
Headquarters
San Francisco, California, United States
Team size
1-10 employees
Industry
Software Development
Sub-industry
Document AI / AI infrastructure
Offices
San Francisco, California, United States
Business model
usage-based API
self-serve SaaS
enterprise contracts
Funding
Stage
pre-seed / early-stage VC
Total raised
$500K
Latest round
Convertible Note
Latest amount
$500K
Aug 2025
Jan 2025 · Y Combinator
Y Combinator
People & investors
Work & culture
Accuracy-first, benchmark-driven, and engineering-heavy; the company presents itself as a fast-moving infrastructure startup focused on replacing fragile document pipelines with reliable, schema-validated parsing for regulated enterprise use cases.
Compensation
YC job listings show founding roles around $120K-$300K base pay plus roughly 0.10% to 1.00% equity; no broader public employee compensation bands were found.
Benefits
usage-based pricing
annual billing discounts
SSO/SAML
zero data retention
HIPAA BAA
EU/US data residency
priority queue
white-glove onboarding
custom SLAs
VPC/on-prem/air-gapped deployment
SOC 2 Type II report
named TAM
24x7 support
dedicated Slack channel
Product & market
Pricing
Usage-based per-page pricing with standard and growth tiers plus custom enterprise pricing for regulated/high-volume deployments.
Differentiators
- —Dual-stream vision architecture that reads both content and layout
- —Schema-validated JSON and Markdown outputs
- —Confidence scores and bounding boxes on extracted elements
- —Supports text, tables, charts, forms, handwriting, and mixed layouts
- —Enterprise deployment options including VPC, on-prem, and air-gapped
- —Public benchmark leadership claims, including #1 on olmOCR-Bench
- —Document parsing, extraction, splitting, and classification in one API
- —Designed for finance, legal, and healthcare workflows
Technology
dual-stream vision models
computer vision
OCR
multimodal models
Python
TypeScript
Docker
Kubernetes
Terraform
Redis
GitHub Actions
PyTorch
TensorFlow
MLflow
React
REST API
MCP server / Claude tool-use schemas
AWS
Azure
GCP
Customers
Fortune 150 banks
NASDAQ-listed enterprises
10+ YC startups
finance teams
legal teams
healthcare teams
Competitors
LlamaIndex / LlamaParse
Unstructured
Reducto
Landing AI
Azure Document Intelligence
AWS Textract
Gemini
Mistral
Scale & contact
Estimated revenue
≈$550K ARR (estimated)
Latest news
Startuply.vc · May 2026
Profile/article highlighting the startup's enterprise focus, estimated $550K revenue, and competitive positioning versus open-source and commercial document-AI tools.
Y Combinator · Feb 2026
YC launch post describing the product, customer wedge, founder background, and claims about processing millions of pages for Fortune 150 banks, NASDAQ-listed companies, and YC startups.
Fondo · Feb 2026
Third-party launch coverage reiterating the product positioning, enterprise customer claims, and benchmark comparisons against LlamaIndex, Gemini, Mistral, and Unstructured.
Jobs and careers at Unsiloed AI
0 jobs at Unsiloed AI
No open positions
Check back later for new openings