Pulse

A systems-level look at how regulated enterprises evaluate document AI beyond accuracy, with emphasis on consistency, structure, and trust in production. | Pulse AI

Advanced spreadsheet parsing is now generally available in Pulse, following millions of XLSX pages processed in production and major improvements in how complex, multi-tab spreadsheets are represented and understood at scale. | Pulse AI

Pulse introduces new models that parse complex spreadsheet structures with superior accuracy. | Pulse AI

A clear, practical overview of agentic OCR, explaining how adaptive reasoning improves document extraction at scale and how hybrid architectures balance flexibility, accuracy, and production reliability. | Pulse AI

Pulse builds AI-powered document processing solutions for finance, healthcare, insurance, and other sectors.

Self Serve is free up to 20k credits with no page limits. Enterprise adds custom volume, HIPAA BAA, zero-data-retention, and on-prem or VPC deployment.

The service processes complex PDFs and scans into verified structured data and states compliance with SOC 2, HIPAA, and GDPR.

Pulse Software's privacy policy describes its data collection, use, and protection practices, GDPR compliance, data security, and privacy rights.

The page outlines the terms of service for Pulse Software, Corp. cloud services, including subscription terms, pricing, payment processes, and the legal framework.

Pulse launched Selection Mark, a purpose-built model for detecting handwritten checkmarks, circles, underlines, and strikethroughs on forms. It adds selection state to Pulse’s standard extraction API, with processing taking under 20 milliseconds per page.

Jul 2026 · Pulse BlogNews article

Pulse launched its Model Context Protocol (MCP) server, making eight tools available to MCP clients. The server exposes extract, apply_schema, generate_schema, split_document, extract_tables, batch_extract, run_pipeline, and get_job.

Jun 2026 · Pulse BlogNews article

Pulse released PulseBench-Tab, an open-source benchmark for multilingual table parsing and extraction. It contains 1,820 human-annotated tables in nine languages and introduces a novel scoring metric called T-LAG.

Apr 2026 · Pulse BlogNews article

Pulse, a startup specializing in unstructured data preparation for machine learning models, raises $3.9 million in a funding round led by Nat Friedman and Daniel Gross. The company aims to address AI’s unstructured-data bottleneck by preparing that data for machine learning models.

Feb 2025 · Business InsiderNews article