About — Roysa

Our Mission

Roysa makes AI for unstructured data traceable. That's the whole point — and why we named our product TrueTrace.

We believe AI should explain its reasoning, not just give answers. When organizations process critical material — legal contracts, government records, site photos, recorded interviews, training videos — they need to trust the results. That trust can only come from tracing every answer back to its exact source.

Our approach combines state-of-the-art multimodal AI — both text detection and object detection — with a fundamental commitment to traceability. Every answer points to the exact word on a page, the object in a photo or video frame, or the moment in an audio recording. Every result is verifiable.

🎯 Transparency

✓ Verifiability

🛡️ Trust

The Problem We Solve

AI answers that can't be traced back to their source can't be trusted.

Most AI tools stop at "this answer is somewhere on page 3." They don't tell you which word, don't recognize objects in your images, and don't work on audio or video at all — so there's no way to verify what they say.

✕ Answers you can't trace back to their source

✕ Only text detection — ignores objects in images and video

✕ Doesn't work on audio or video at all

✕ Hallucinates or misreads values with no way to verify

Our Approach

Every answer must be traced back to its exact source.

Traceability is the product. We don't just tell you what we found — we show you the exact word on the page, the object detected in the photo or video frame, or the moment in the recording where we found it, how confident we are, and how you can verify it yourself. Text detection and object detection, side by side, every result traceable. That's TrueTrace.

Key Principles

No Black-Box Answers

Every answer is transparent. See the reasoning, not just the result.

Evidence-First AI

Visual grounding means every value points back to its source on the document.

Designed for Review

Built from the ground up to support human verification and audit requirements.

Real-World Ready

Engineered to handle noisy scans, skewed images, and imperfect documents.

Our Story

Roysa was born from a simple observation: organizations are drowning in documents, photos, recordings, and videos — and existing tools either read only text, only find rough sections, or don't show where their answers came from.

We saw teams spending countless hours manually verifying AI outputs because the AI pointed at a vague paragraph instead of the exact word. We saw tools that ignored objects in images and couldn't touch audio or video at all.

We built Roysa to fix this — one assistant that reads every word, recognizes objects, listens to recordings, watches videos, and pins every answer to its exact source.

Founder

Ahmad Shirazi

Ahmad Shirazi is the founder of Roysa. He holds a Ph.D. in Computer Science and Engineering from The Ohio State University, where his research focused on document, audio, and video understanding. His work has been published in leading machine learning venues and includes answer localization in documents, visual question answering, and efficient multimodal architectures.

He founded Roysa to bring advanced AI research into real-world document, audio, and video workflows where accuracy and verifiability matter.

Roysa is a member of NVIDIA Inception

Roysa is a member of the NVIDIA Inception program, which nurtures dedicated and exceptional startups who are revolutionizing industries with advances in AI and data science.

This partnership gives us access to NVIDIA's cutting-edge GPU technology, deep learning expertise, and go-to-market support — helping us push the boundaries of word-level, object-aware AI across documents, images, audio, and video.

Security & Privacy

Clear, customer-first data handling

No Training on Customer Data

Your documents, audio, and video are never used to train our models. We maintain a strict separation between customer data and model development.

About Roysa