amd Applications

OmniDoc — Talk to Any Document

Upload complex documents — PDFs with charts, tables, scans — and have a natural languageconversation about everything, including the visuals.

Boundary Forge

Boundary Forge helps enterprises safely deploy any LLM by automatically finding model failure boundaries, benchmarking risk, and converting failures into runtime guardrails. Demonstrated with Qwen 2.5-72B on AMD Cloud.

OrdoFlow — AI Quote-to-Cash Copilot (AMD Track 1)

AI Quote-to-Cash Copilot turns quote context into prioritized next actions, risk flags, and customer-ready messaging inside OrdoFlow’s live quote workflow.

Thor v2 — RAG-Free Fitness Intelligence

Thor v2 fine-tunes Qwen3-8B on AMD MI300X to answer fitness questions from weights alone. No RAG. Citation keys are emitted inline and validated against a locked registry. 7,118 training examples. 14 population profiles. 100% JSON contract pass rate.

Agronomist

Resume Compact

An intelligent HR analysis system powered by FastAPI and local LLMs that automates resume screening and candidate evaluation.

Swarm

A multi-agent collective intelligence for DevOps orchestration and automation, powered by the GPUs of the AMD Developer Cloud.

TempoGraph: Local Multimodal Video Analysis

A fully-local multimodal video analysis pipeline that transforms raw video into structured entities, events, and timelines using YOLO26, Whisper.cpp, Depth Anything V2, and Qwen3.5-VL — all running on consumer AMD hardware. No cloud, no API keys.

Embodied Red-Team Arena

“A future home robot, drone, warehouse bot, or inspection agent may follow user instructions correctly but still behave dangerously in the physical world. RoboHazard Arena fin”

Ghostchain AI+SafeFi Omega

Ghostchain AI+SafeFi OMEGA (Auth ID: 287F-930E) is a sovereign AI entity that prevents neural drift by gating trading logic through thermal sensors and utilizing the x402 protocol for secure machine-native payments

AMD Inference Gateway

ROCm-native adaptive inference gateway with MI300X-aware dynamic batching

ROCmPort AI

ROCmPort AI is a 5-agent pipeline that migrates CUDA kernels to AMD ROCm/HIP. It detects hardware-specific bugs hipify misses, compiles with hipcc, profiles with rocprof on real MI300X hardware, and iterates until the output is correct and optimized.

SentinelOps: AI Autonomous Incident Response

SentinelOps is a multi-agent system that detects, investigates, and generates responses to cybersecurity incidents in real time. It reduces response time from hours to seconds through a cascade of 4 specialized AI agents running on AMD hardware.

AssemblyMind

A real-time multimodal agent that reads your schematic and watches your workbench camera to catch wiring errors, reversed parts, and bad connections before power-on. Powered by AMD MI300X.

Urban Expansion Detector

Fine-tuned Qwen2.5-VL-72B on AMD MI300X to detect urban expansion from Sentinel-2 satellite imagery. Identifies built areas, estimates urban coverage fraction, and generates annotated corridor reports for transit infrastructure planning in India.

AEGIS

An autonomous, multi-agent clinical operating system that monitors patient telemetry, analyzes medical imagery, and prevents fatal drug errors in real-time, designed for secure, on-premise AMD deployment.

ROCmPort AI: Agentic Migration Suite

A 3-agent CrewAI pipeline powered by Qwen3 that automates scanning, patching, and benchmarking legacy PyTorch/CUDA repositories for native execution on AMD Instinct MI300X.

MediAgent — Autonomous Radiology Pipeline

MediAgent is a 5-agent AI radiology assistant running 100% locally on AMD MI300X. Upload any X-ray, MRI, or CT scan and receive a full clinical report with differential diagnoses, confidence scores, and QA validation in under 90 seconds.

PaperHawk: Audit-Grade Document Intelligence

Multi-agent document intelligence that catches what RAG misses. 14 deterministic domain checks, 5+1 anti-hallucination layers, and a 5-tool agentic chat — running Qwen 2.5 on AMD Instinct MI300X via vLLM. Open source, MIT licensed.

APIlot: Intelligent API Navigator

APIlot is an AI-powered navigator that eliminates endless documentation scrolling. Drop an API URL, and it instantly extracts every endpoint and required parameter into a dynamic UI for rapid payload generation.

Raksha Setu: AI emergency bridge

Cardiac arrest causes brain damage in 4 minutes. Ambulances take 18. Raksha Setu is AI-powered emergency mesh that closes 14-minute gap, dispatching trained volunteers on site before ambulance arrives, guided in real time by Qwen running on AMD MI300X.

Code Guardian

Autonomous 5-agent AI pipeline — scans any GitHub repo for bugs, patches them, writes tests, runs them in an isolated sandbox, and opens a PR. AMD MI300X + Llama 3.3 70B.

WarpShift — CUDA to ROCm AI Migration Agent

AI agent that migrates CUDA repositories to AMD ROCm in 90 seconds. Combines HIPIFY with Google Gemini 2.5 Flash to detect what deterministic tools miss. Validated live on AMD Instinct MI300X, 5.5x faster than CUDA baseline.

AI-Documentation-Agent

An intelligent agent that auto-generates technical documentation from code and context, powered by AMD cloud infrastructure.

Hands for AI

An open-source agentic engine that drives Chrome through every available channel — DOM, accessibility tree, network interception, JS evaluation, multi-tab and multi-context. Each task graduates to a direct API call after first run. Linux + ROCm on AMD.

The Architect’s Eye

The Architect’s Eye is an autonomous, multi-agent AI system for construction safety. It leverages multimodal vision and regulatory auditing to detect site hazards in real-time, automatically generating precise, OSHA-compliant reports to prevent accidents.

VICTOR - AI ER Triage

Voice-first AI triage agent that catches atypical cardiovascular presentations standard ER triage misses, especially in underserved populations.

BizAI — Multi-Step AI Agent on AMD MI300X

A Streamlit AI agent that automates business emails and invoices through a visible 3-step reasoning loop. Mistral 7B runs on a single AMD Instinct MI300X via vLLM + ROCm at ~223 tokens/sec. Open-source end to end, no vendor lock-in.

PatchPilot AI

PatchPilot AI is an agentic debugging copilot for C++ game engine developers. It reviews multiplayer/networking source code, detects bug and performance risks, suggests fixes, and generates practical tests through a multi-agent

REPOMIND

Open-source repo-scale coding agent verified on AMD MI300X. Single GPU, Qwen3-Coder-Next-FP8, 256K context. 31/31 users at 8K-64K, 3/3 needle 200K, 9/9 correct on pytorch/vision (1.3M tokens). 124-min stress test, AITER tuning A/B. MIT.

Vision-Link AI Healthcare Project

A multilingual AI-powered healthcare dashboard bridging diagnostic gaps in Africa. We use AMD ROCm and Google Gemini to provide real-time vitals tracking and medical insights in local dialects (Hausa, Yoruba, Igbo).`

tests-tests

tests-tests tests-tests tests-tests tests-tests tests-tests tests-tests tests-tests tests-tests tests-tests tests-tests tests-tests tests-tests tests-tests tests-tests tests-tests tests-tests tests-tests te

RustVital-AMD

Zero-trust medical AI triage gateway in pure Rust. PII redaction before AI, Filecoin CID + Base L2 on‑chain audit. Ready for AMD MI300X.

SimAgent — AI Engineering Simulation Bot

Al Agent Orchestrator for Real-Time Engineering Simulation. Powering the next generation of CAD/CAE workflows on AMD ROCm™ hardware.

GHOST Chimera

Ghost Chimera is a local-first agent orchestration project.

ROCmPilot

ROCmPilot is a multi-agent AI system that scans CUDA-first GitHub repos and generates ROCm migration findings, patch previews, validation plans, and reports.

Enziu

ENZIU audits your actual insurance policy PDF and returns a scored ENZIU Index based on Clarity, Coverage, and Claim Efficiency, with red flags, page citations, and a Deep Dive Q&A. No accounts. No stored data.

InsightAgent — Multi-Agent AI on AMD MI300X

Watch AI agents collaborate in real-time on AMD MI300X GPUs. Upload any CSV, and 4 specialized agents — Parser, Vision, Analyst, Reporter — analyze your data, generate executive reports, and let you chat with your data using natural language.