LightOn

One API to run them all. Agents do not ask nicely. They dump raw PDFs, garbled tables, and off-domain questions into the same thread. 📄Parsing: LightOnOCR-2 turns scans, tables, handwriting, and multi-column layouts into structured output. 20+ languages natively. Send the raw file. No Unstructured, no LayoutLM, no glue code. 🧠Extraction: Pull any key-value, entity, or field you care about straight from documents and images. Define the schema, get structured JSON back. 🔍Search: Dense embeddings, sparse lexical terms, and NextPlaid late-interaction vectors, powered by LateOn, which generalizes across domains with no fine-tuning. One query hits all three signals. The index picks the signal; you do not pick the index. One API key. One dashboard. One spend cap. Three endpoints. Zero pipeline maintenance. claim your access : console.lighton.ai

@LightOnIO × @Linkup_platform On May 20 in Paris, @LightOnIO and @Linkup_platform are bringing together the AI community for an evening dedicated to Search, Agents, and LLMs, hosted by @ParisMLgroup . A series of technical talks will explore why search is becoming a core layer of enterprise AI and how the next generation of AI systems is being built. 📍 Paris 📅May 20, 2026 ⚠️ Limited spots ! Register here: meetup.com/paris-machine-…

One API to feed any model, power any agent. One Playground, three endpoints, zero config. Request Early Access!

Un fournisseur dépose le bilan à 9h47. Vous recevez une alerte : « plus de 1 M€ d’exposition, call avec la direction à midi." Ce cas d'usage explore la capacité d’une IA à raisonner à partir de documents d'enterprise fragmentés. Quand “avoir à peu près raison” revient à entrer en réunion avec le mauvais chiffre. 📰 Lire l’analyse complète ici : lighton.ai/fr-blog-posts/… 💻Testez le scénario sur votre propre corpus documentaire → console.lighton.ai

Reason-ModernColBERT topped BrowseComp-Plus with just 149M parameters. Now, Agent-ModernColBERT adds ~10% on top. Reaches GPT-5 + Qwen3-8B with GPT-OSS-120B. Still 149M parameters. Fully Open. Smaller. Cheaper. Kudos to @antoine_chaffin for the work 👏🏻 Full benchmarks, methodology, model, data, and training code in the blog ↓ lighton.ai/lighton-blogs/…

One API to feed any model, power any agent. One Playground, three endpoints, zero config.

@LightOnIO × @Dassault3DS Some conversations signal where the market is heading. On June 9, LightOn will be part of @outscale EXPERIENCES 2026 to discuss a shared vision of enterprise AI: sovereign, secure, and built for operational deployment at scale. As AI moves into critical environments, infrastructure, governance, and execution can no longer be separated. More soon. 📍 CNIT Forest, Paris La Défense hashtag#SovereignAI hashtag#EnterpriseAI hashtag#OUTSCALEExperiences hashtag#LightOn hashtag#DassaultSystemes

Retrieval is all about signal, architecture, not raw power. 🦾The late interaction LateOn 0.1B competing with the dense Gemini 2

RT @IgorCarron: Tiny @LightOnIO 's LateOn on the same level as GPT-5 Query Rewriter! Onward to beating the Oracle with small and Open Sour…

Better signal. Faster inference. XTR training and WARP index are now merged into PyLate: Same late-interaction signal. Lower inference cost at scale. The Late Interaction train is not slowing down. Jump in!

From deploying agents → to owning your data, models, and infrastructure. That shift is being explored today at @gosimfoundation Paris 2026, a key European event for open-source and sovereign AI. 🎤Today, @staghado took the stage for @LightOnIO : “LightOnOCR: Pushing the Performance–Efficiency Pareto Frontier of Open OCR Models” 🔍 Parsing is the first step of any AI system Before anything else, your data needs to be extracted, structured, and usable. 👉🏻 Discover LightOn orchestrated pipeline, from parsing to grounded answers: console.lighton.ai

The plumbing era of RAG is over. Parsing. Chunking. Embeddings. Reranking. Search. Five moving parts, one fragile seam between each, and a roadmap quietly disappearing into the gaps. The teams shipping fastest have moved up the stack: 🎯 Domain expertise : the part a competitor cannot copy ⚡️ Product velocity : features instead of chunk-size debugging Retrieval infrastructure is solved. The moat is what you build on top of it. Full post → lighton.ai/lighton-blogs/… More from us on this, very soon.

The data you need to benchmark enterprise RAG is the data no one shares. So we built it. Today we're releasing EDiTh on @huggingface , an open benchmark for enterprise retrieval, designed around the questions executives actually ask. What's inside: 📄 1,004 PDFs across 6 languages and 3 formats 🎯 36 use cases with full answer keys 🎭5.5% plausible distractors (the kind that fool real systems) 🏢 All built around Véracier Industries, a fictional €1.8B industrial group Each use case mirrors a real executive question, the kind that usually takes a team, a week, and a stack of PDFs to answer. If you're building, buying, or evaluating RAG for the enterprise, EDiTh gives you something the public benchmarks don't: realistic documents, realistic noise, realistic stakes. Built by Adèle Guignochau and @IgorCarron at @LightOnIO . Read the release 👉🏻 lighton.ai/lighton-blogs/… Dataset on Hugging Face 👉🏻 huggingface.co/datasets/light… #EnterpriseAI #RAG #SovereignAI 📄

One API to run them all. Agents do not ask nicely. They dump raw PDFs, garbled tables, and off-domain questions into the same thread. 📄Parsing: LightOnOCR-2 turns scans, tables, handwriting, and multi-column layouts into structured output. 20+ languages natively. Send the raw file. No Unstructured, no LayoutLM, no glue code. 🧠Extraction: Pull any key-value, entity, or field you care about straight from documents and images. Define the schema, get structured JSON back. 🔍Search: Dense embeddings, sparse lexical terms, and NextPlaid late-interaction vectors, powered by LateOn, which generalizes across domains with no fine-tuning. One query hits all three signals. The index picks the signal; you do not pick the index. One API key. One dashboard. One spend cap. Three endpoints. Zero pipeline maintenance. claim your access : console.lighton.ai

Retrieval is solved. One API to feed any model, power any agent. Three endpoints, zero config: /parse /extract /search LightOn Console dropping soon. Sandbox access, ship as you sign up! 🔗 console.lighton.ai