Compute-aware HPC policy and resource governance engine for OmniBioAI — enforces GPU/CPU quota limits, Slurm-aware scheduling policies, cluster access control, and zero-trust execution decisions. Prevents resource monopolization on shared HPC infrastructure and ensures fair, auditable compute allocation across research teams.
Electron + React + Vite desktop app for OmniBioAI — provides a visual workflow builder, plugin launcher, and unified control plane for local, HPC, and cloud bioinformatics execution. Connects to the Django backend via IPC and REST, integrates Docker runtime and local LLM inference via Ollama.
Shared React component library for OmniBioAI — Button, Badge, Card, Input, StatusDot, and Spinner built on @man4ish/design-tokens. TypeScript-first, tree-shakeable ES+CJS build via Vite, published to GitHub Packages. Zero hardcoded colors; all styling delegated to the token layer.
Full GPU-accelerated AI development environment for OmniBioAI — built on NVIDIA PyTorch 25.10 with CUDA, includes R, MySQL, JupyterLab, Hugging Face Transformers, and Ollama. Designed for bioinformatics ML workflows requiring both deep learning (PyTorch/CUDA) and statistical genomics (R/Bioconductor) in a single reproducible container.
Runtime data directory for the OmniBioAI platform — directory structure, sample bioinformatics files, and registry configuration for bootstrapping a local instance.
FastAPI-based ToolServer for OmniBioAI — provides validated, asynchronous execution APIs for bioinformatics tools. Accepts tool invocation requests, validates inputs against declared schemas, executes tools in isolation, and returns structured results. Acts as the execution boundary between the platform and individual tools.
Distributed security audit and observability system for OmniBioAI — Redis Streams-based real-time event pipeline capturing authentication events, policy decisions, data access, and system actions across all platform services. Provides zero-trust traceability, tamper-evident audit logs, and scalable event replay for compliance and forensics.
Production-grade FastAPI authentication service for OmniBioAI — JWT-based login with access and refresh tokens, role-based access control (RBAC), MySQL-backed user and role management, OAuth2 password flow, and token revocation. Central identity provider for all platform services.
ARM64-compatible Docker and Singularity images for bioinformatics and ML tools — FastQC, STAR, BWA, GATK, DESeq2, PyTorch, and more. Built for NVIDIA DGX Spark (aarch64) and deployable via Slurm. Declarative YAML-driven image definitions; part of the OmniBioAI reproducible tool execution stack.
Official Python SDK for OmniBioAI — typed client for the object registry API (list, filter, paginate, fetch by ID), notebook launch integration for JupyterLab and RStudio, and authentication helpers. Built with Hatchling, 95% test coverage enforced, supports OMNIBIOAI_BASE_URL and OMNIBIOAI_TOKEN environment-based configuration.
Multi-repo AI intelligence hub for OmniBioAI — hybrid retrieval system combining vector search, knowledge graph traversal, and plugin-aware retrieval over the full codebase and biomedical corpus. Streaming LLM responses with a React visualization UI for code navigation, cross-repo dependency analysis, and biomedical Q&A.
Unified design token system for OmniBioAI — 47 CSS custom properties (colors, typography, spacing, radius, shadows) shared across all frontend surfaces. Zero hardcoded values; swap themes by overriding tokens. Ships as CSS, JS, and TypeScript with no build step.
MySQL initialization scripts for the OmniBioAI platform — creates databases and loads schema/seed data on first container startup.
Zero-trust API gateway for OmniBioAI — handles authentication (JWT verification), policy enforcement (RBAC/ABAC), request routing to downstream services, rate limiting, and secure service-to-service communication. All inbound traffic to the platform passes through this gateway; no service is directly addressable without authorization.
Production-grade ML model registry for OmniBioAI — versioned storage, SHA-256 integrity verification, provenance tracking, and full lifecycle management (register, promote, deprecate, retire) for AI/ML models across local disk, cloud object storage, and HPC filesystems. Supports audit trails and model lineage queries.
Minimal cloud-agnostic container execution runtime for the OmniBioAI Tool Execution Service — enforces a strict input/output contract for tools running on AWS Batch, Azure Batch, and Kubernetes. Handles environment injection, output collection, and exit-code-based status reporting with no cloud-specific dependencies in tool images.
Unified zero-trust security SDK for the OmniBioAI ecosystem — provides IAM-based authentication, service-to-service (S2S) mutual auth, policy enforcement client, Redis-backed token caching, and structured audit log emission. Used by all OmniBioAI services to enforce consistent security posture without reimplementing auth logic.
Video library and interactive Getting Started guide for OmniBioAI Studio — MP4s served via nginx with zero backend dependencies. Covers installation, first run, plugin usage, workflow building, and agent orchestration. Videos stored in Git LFS; container deployable with a single docker run command.
FastAPI-based control plane for OmniBioAI — provides HTTP/TCP/disk health monitoring across all platform services, live operational dashboards, and ecosystem-wide report generation (architecture diagrams, codebase stats, test coverage, system health summaries). Single pane of glass for platform operators.
To associate your repository with the omnibioai topic, visit your repo's landing page and select "manage topics."