DocSlight
Parse PDFs, scans, images, and Office files into structured Markdown and JSON for insurance workflows.
Turn complex claims, policies, and medical records into structured data using local or cloud engines.
DocSlight is an open-source document parsing and data extraction engine designed to convert unstructured files like PDFs, scans, images, and Office documents into structured JSON or Markdown. It is built for insurance professionals handling heavy documentation workflows, policy analyses, and automated claims processing.
What it is
The platform combines OCR supporting 80+ languages with automated document layout analysis and structured field extraction. You can deploy it locally or via a cloud API, and interact with it through a Python SDK, CLI, or a Docker-based Web UI.
Who it's for
It suits insurance technologists and developers building internal RAG pipelines, AI agents, or automated document ingestion systems for claims and underwriting. It does not suit end-users expecting a ready-to-use desktop application without technical deployment.
Key features
Multilingual OCR Support
Recognizes text across 80+ languages with automatic language detection for global policy documents.
Structured Data Outputs
Converts complex document layouts directly into Markdown or JSON formats for downstream pipelines.
Flexible Deployment Modes
Run entirely locally via Python or Docker, or switch to the cloud mode for higher accuracy extraction.
Multiple Interfaces
Interact with the engine using a web UI, a command-line interface, or a Python software development kit.
Up and running in 3 steps
Our honest verdict
What's great
- Supports legacy Office formats and image files alongside standard PDFs.
- Provides both a free local mode and a cloud API option.
- Outputs structured JSON with bounding-box traceability in enterprise editions.
- Includes a Docker-compose configuration for quick browser-based Web UI testing.
Where it falls short
- Local mode requires a local Large Language Model for advanced PDF and image extraction.
- Running the Web UI locally requires setting up Docker and Docker Compose.
- Advanced features like human review workflows and auto-classification require enterprise editions.
Pricing & license
Free under its LGPL-3.0 open-source licence for local use, with optional cloud API modes that offer free quotas alongside paid enterprise tiers.
Open remote AI-training roles
880 open remote AI-training roles, updated November 15, 2026; most pay $20–$175 an hour. Zekai lists them free; the application goes to the employer’s platform.
Alternatives
Use it, or have us set it up
FAQ
Can I run DocSlight completely offline?
Yes, you can use local mode without registration, though local LLMs may be required for certain extraction tasks.
What file formats are supported?
It supports PDFs, images, scans, and Office files like Word, PPT, and Excel.
