Pulse

Pulse is a document intelligence platform that converts PDFs, scans, images, spreadsheets, and Office files into structured data. It addresses the difficulty of extracting reliable content from complex layouts, multi-column pages, tables, charts, and irregular document formats. The processing pipeline combines layout detection with component detection models, low-latency OCR, reading order algorithms, table structure recognition and parsing, and fine-tuned vision language models for converting charts, tables, and figures. Additional capabilities include Classify for document type identification and routing, Agentic Chart Reconstruction, document metadata retrieval, bounding-box coordinates, webhook callbacks, multilingual OCR, and structured JSON output with user-defined schemas.

Pulse is aimed at enterprise and AI-native teams, with deployments available in private VPC, on-prem, air-gapped, and multi-cloud or hybrid configurations. It supports SAML/OIDC SSO, granular RBAC, zero-data-retention, regional data residency, and HIPAA BAAs. Packaging includes a free Self Serve plan with included credits and credit-based overage, a Pro tier, and a custom Enterprise tier with custom invoicing, upgraded rate limits, and custom fine-tuning. A limited-usage sandbox API key is available on request.