RightNow AI
RightNow AI is a YC-backed GPU research lab focused on model-hardware co-design, GPU kernel optimization, and AI infrastructure. Its products aim to let developers and teams own more of their AI stack rather than depend on closed-source APIs.
At a Glance
- AI developers and GPU programmers
- Machine-learning researchers and infrastructure teams
- Startups and enterprises building AI agents
- Teams serving open-source language and multimodal models
- +2 more
AI Tools by RightNow AI
(3)
RunInfra
Hosted Open Model API Gateway
OpenFang
Open Source Agent OS in Rust
RightNow AI
GPU Kernel Editor with AI
Discussions
No discussions yet
Be the first to start a discussion about RightNow AI
Latest News
RunInfra published 'With software alone, one B200 beats the LPU and gets close to Cerebras.'
RunInfra published '$0.09 and $290.12: What Actually Moves Your Inference Bill.'
RunInfra published measurements on serving Kimi K3 with vLLM.
RunInfra published a vLLM vs. SGLang vs. TensorRT-LLM serving benchmark.
Products & Services
A multimodal decision API/model that accepts text, files, codebases, PDFs, images, voice, and video in one call and returns answers with probabilities. The public system-zero model is always the latest version.
Hosted open-model inference for agents, with one API key, session-aware warm-replica routing, cached-context pricing, and OpenAI-compatible Chat Completions plus Anthropic-compatible Messages APIs.
GPU-native development environment for CUDA and related GPU languages, combining profiling, benchmarking, AI optimization, GPU virtualization, and GPU emulation.
Open-source research/software project that profiles PyTorch models, tests Triton variants, and uses measured bottlenecks for autonomous GPU-kernel optimization.
Market Position
RightNow positions itself at the intersection of GPU-native developer tooling, kernel optimization research, and open-model inference. RunInfra is marketed as a lower-cost, agent-oriented alternative to closed-source model APIs, while System Zero is benchmarked publicly against OpenAI's GPT-6 Luna and GPT-6 Astra on speed, cost, and decision accuracy; the older editor product differentiates through integrated profiling, emulation, and optimization rather than a general-purpose coding workflow.
Leadership
Founders
Jaber Jaber
Technical co-founder. YC identifies him as previously at Microsoft and Nujoom AI; his public profile also lists an AI bachelor's degree at Al Hussein Technical University and Stanford machine-learning specialization.
Osama Jaber
Commercial co-founder/CMO. Previously founded Antilab, an AI healthcare/diagnostic startup described as helping reduce antibiotic misuse; public profiles identify him as an Al Hussein Technical University alumnus.
Executive Team
Jaber Jaber
Technical Co-founder
Previously at Microsoft and Nujoom AI according to YC; AI/data-science background at Al Hussein Technical University.
Osama Jaber
Commercial Co-founder / CMO
Founder of Antilab, an AI healthcare/diagnostic startup; leads commercial and marketing work at RightNow.
Board of Directors
Founding Story
RightNow began as a CUDA-focused code editor intended to make GPU development and optimization more accessible. The company expanded that initial vision into a research lab and an inference platform designed around model-hardware co-design and open models.
Business Model
Revenue Model
RunInfra combines prepaid, pay-as-you-go per-token Model API usage with monthly coding-agent plans and custom-priced enterprise contracts for dedicated infrastructure, compliance, volume, and SLAs. System Zero is offered as a pay-as-you-go API with usage-based token/question pricing.
Pricing Tiers
Prepaid credits; pay per token at published model prices; credits last a year; no subscription or per-seat fee.
Coding plan with included usage that resets every five hours and weekly.
Coding plan with included usage that resets every five hours and weekly.
Coding plan with included usage that resets every five hours and weekly.
Dedicated infrastructure, custom volume/terms, audit logs, RBAC, SOC 2 Type II, up to 99.99% SLA, dedicated CSM, and private Slack.
Target Markets
- AI developers and GPU programmers
- Machine-learning researchers and infrastructure teams
- Startups and enterprises building AI agents
- Teams serving open-source language and multimodal models
- Organizations seeking dedicated, compliant inference infrastructure
- Edge-AI and memory-constrained device developers
- Developing, profiling, and optimizing CUDA and GPU kernels
- Serving open-source language models for coding and autonomous agents
- Building AI-agent applications with OpenAI or Anthropic SDK compatibility
- Low-cost, low-latency production inference
- Document AI and retrieval/search applications
- Edge inference on memory-constrained devices
- NVIDIA
- Hugging Face
- Runway
- Together