LightOn

LightOn is an AI search and document intelligence platform for parsing files, extracting structured data, and finding grounded answers with citations.

AI Search Engine AI Knowledge Base AI Document Extraction AI OCR

Visit Website

Overview

LightOn is a document intelligence and retrieval platform built around three API endpoints: `/parse`, `/extract`, and `/search`. It is designed to turn files into structured output, extract specific fields with guided schemas, and return grounded search results with citations.

The product is positioned for developers, teams shipping production systems, and large organizations that need search and reasoning over their own documents. The site emphasizes OCR for scans and complex layouts, JSON extraction, hybrid retrieval, source-backed answers, and deployment options that include EU sovereign hosting and enterprise environments.

Core capabilities

OCR and document parsing

Parse scanned documents, tables, handwriting, and multi-column layouts into structured Markdown using the `/parse` endpoint.

Schema-guided extraction

Extract fields, entities, or key-value pairs by providing a JSON schema, with structured JSON returned in response.

Hybrid search with citations

Run grounded retrieval with dense, sparse, and late-interaction signals, and return source passages with each result.

Browser-based playground

Test endpoints in a browser Console, then copy the code into your project without installing extra tooling.

Scoped access and governance

Use workspaces and chunk-level ACLs to scope access for teams and agents, with mirrored source permissions and audit logs on paid plans.

Flexible deployment and pricing

Choose from usage-based API plans or enterprise deployment options, including cloud, VPC, on-prem, and regional hosting.

Agent-friendly integration

Connect through MCP-native workflows and integrate the API with agents or applications that already speak Model Context Protocol.

Common use cases

Prepare documents for automation
Use `/parse` to convert scanned PDFs, tables, handwriting, and multi-column documents into structured text before downstream processing.
Capture specific fields from documents
Use `/extract` when you need invoice numbers, contract clauses, claim IDs, or similar fields returned in JSON that matches your schema.
Build grounded document Q&A
Use `/search` to answer questions from company files with cited passages so users can verify the source of each answer.
Support agentic reasoning workflows
Use the API inside agent workflows where the system needs to retrieve relevant passages, chain steps, and produce answers based on internal documents.
Search sensitive internal knowledge
Use the enterprise interface and deployment options when teams need secure access to sensitive knowledge across workspaces and permissions.

Pros and Cons

Pros

Covers the full document workflow from parsing to extraction to cited retrieval.
Returns structured outputs that are useful for automation, not just text responses.
Supports browser testing in Console before implementation.
Offers usage-based pricing with a free Starter tier and higher-touch Business and Enterprise options.
Provides deployment choices for organizations with sovereignty or infrastructure requirements.

Cons

The public pages do not provide a complete connector list, so integrations beyond the named examples are unclear from the source provided.
Some capabilities are described at a high level without detailed limits, implementation steps, or output examples for every workflow.
Enterprise deployment, security, and connector details require consultation rather than being fully specified on the public pages.

FAQ

What does LightOn’s API do?

LightOn provides three main API endpoints: `/parse` for OCR and document structuring, `/extract` for guided JSON extraction, and `/search` for grounded retrieval with citations.

Does LightOn offer a free tier or paid plans?

Yes. The pricing page shows a Starter tier with free, usage-based access, a Business tier for teams, and an Enterprise tier with tailored deployment and support.

How is LightOn deployed and tested?

The pricing and API pages mention browser-based testing in Console, copy-paste code examples, and deployment options including EU sovereign hosting, VPC, on-prem, and regional options for Enterprise.

How does LightOn handle citations and access control?

LightOn says answers ship with source passages and retrieval is auditable, with access control enforced at the chunk level and workspaces used to scope corpora.

Quick Facts

Category: Document intelligence and AI search
Platform: Web app, API, and browser Console
Primary users: Developers, product teams, and enterprises
Source domain: lighton.ai
Key endpoints: `/parse`, `/extract`, `/search`
Pricing model: Free Starter tier, Business plan, and Enterprise contact sales