Explore the interactive games we have built, practical AI experiments we are documenting, and the creative and technical capabilities behind our work. Commercial services are currently paused.
From playable projects to practical experiments, explore verified work rather than a list of claims.
Explore Throttle Outlaws and Circuit Rivals in your browser.
See playable projects → For enthusiasts & developersLearn what worked—and what was slow—on actual hardware.
See technical findings → For collaboratorsShare a research or project enquiry without assuming service availability.
Contact Optizord →Our current browser-based games and an educational development concept. Open a playable project, explore what it involves, or review the research behind our approach.
Move your pointer across a project card to explore its depth. On touch devices, the cards remain steady for easier reading.
The Android version is planned for release on Google Play. The store listing is not live yet; this search link is a placeholder until launch.
Playable gameAn original browser-based driving project. Explore racing action and the current playable build.
Playable gameA developing 3D racing experience exploring driving challenges and interactive environments; features continue to evolve.
Real project promotional graphics supplied by Optizord. Actual gameplay and current builds may differ.
Exploring an interactive way to learn camera angles, cinematic shot composition and AI prompting for animated storytelling. This is a developing concept, not a released game.
Our work brings together browser games, interactive learning experiences, AI-assisted media experimentation and practical software development. We evaluate tools and workflows through hands-on projects rather than promising results before they are proven.
Recent work includes racing games, experiments with AI-assisted production, and an educational animation-directing concept. Explore our active projects and technical findings below.
Exploring AI-enabled workflows, visual storytelling, games, apps and modern web technologies while documenting lessons and practical constraints.
Hands-on engineering experiments with real hardware, AI-assisted creative workflows, connected systems and original games. Outcomes below are reported observations, not independently verified benchmarks or commercial service claims.
Configured ComfyUI on an RTX 4060 gaming PC (16 GB system RAM) for a reported locally executed MiniMax H3 video workflow. Produced 5–15-second clips at a reported direct 1080p output, taking approximately 10–20 minutes per clip. Reference images and video supported consistency across assembled sequences. Also experimented with sound effects, music and voiceovers.
Used Ollama with lightweight Qwen-family models on Intel Pentium dual-core and Core i3 computers without dedicated GPUs. Text generation functioned; one observed response on the Core i3 took approximately a minute and the Pentium took longer. Image, video and audio generation were not practical on those systems.
Explore the implementation methods, practical results, limitations and next verification steps. These are reported first-hand experiments; detailed workflows and independent measurements are not yet public.
Setup: ComfyUI on an NVIDIA RTX 4060 gaming computer with 16 GB system RAM. The experimenter reports executing MiniMax H3 entirely locally, producing 1080p outputs directly without a separate upscale step. This deployment detail has not been independently validated using workflow files or output metadata.
Workflow: Used facial, hand, leg, front, back and side reference images, alongside reference video, to guide character appearance across individual 5–15-second shots. Short shots were assembled into sequences around five minutes long. Audio experiments included SFX, music and voiceovers.
Observed: Approximately 10–20 minutes per generated clip; one concurrent generation at a time. Longer generation attempts sometimes crashed. Character consistency was informally estimated at around 95%, without a standardized measurement method.
Next evidence: Sanitized ComfyUI graph, exact node/model version and execution path, exported 1080p sample metadata, GPU VRAM usage and a documented continuity evaluation.
Setup: Ollama runtime with lightweight Qwen-family text models on Intel Pentium dual-core and Intel Core i3 computers without dedicated graphics cards. Several lightweight model configurations were tried; exact identifiers were not retained for publication.
Observed: Functional local text inference, but slow response generation. An approximate one-minute response was observed on the Core i3; the Pentium took longer. Image, audio and video generation were not practical on these machines.
Next evidence: CPU/RAM specifications, model sizes, quantization, prompt length, first-token latency and measured tokens/second.
Architecture: Created a custom connector arrangement using Python, Node.js and local APIs, with Claude Code assisting the implementation. The experiment investigated communication among local AI models, ChatGPT and Claude to coordinate content-related tasks.
Distribution: Developed automated publishing workflows for Facebook, Instagram and YouTube using APIs and third-party posting technologies. This describes a tested integration approach, not a claim of uninterrupted or fully autonomous operation.
Constraints: AI-service usage limits, orchestration reliability and external platform dependencies affected sustained operation.
Next evidence: A credential-free architecture diagram, example job lifecycle, approval/error-recovery process and publicly shareable logs.
Projects: Throttle Outlaws and Circuit Rivals.
Process: Human-defined game ideas and design direction informed architecture produced with ChatGPT. Claude Code assisted with implementation; the actual builds were evaluated on devices and refined through gameplay testing.
Challenges: Vehicle/game physics required iteration and correction. Character continuity in generative media also needed repeated reference tuning. These were solved through testing and refinement rather than a one-prompt workflow.
Next evidence: Versioned change logs, gameplay comparisons, device compatibility notes and short technical explanations of corrected physics issues.
The outcomes above are experimenter-reported. The ComfyUI workflow path and direct 1080p MiniMax H3 generation claim still need verification using exported node settings and sample media metadata. We have not published repeatable throughput, reliability or standardized quality benchmarks.
Next documentation step: capture a sanitized ComfyUI workflow, model configurations, hardware memory use, generation logs and representative outputs.
These reports describe first-hand experiments, not third-party benchmarks. Further technical specifications and reproducible measurements will be published when available.
A transparent overview of capabilities we have offered before and technologies we are exploring for potential future offerings.
Work areas previously included in our creative, web and digital service portfolio.
AI video advertisements, spokesperson and avatar videos, social media reels, promotional videos, voiceovers, image generation, creative concepts and product mockups.
Business websites, responsive landing pages, AI-assisted static websites and hosting, content updates, website maintenance and technical improvements.
Search-engine optimization, answer-engine and generative-engine discovery work, content strategy, social media campaigns, performance advertising and reporting.
Logo design, social posters, product visuals, marketing creatives, digital branding and campaign-oriented design assets.
Conversational assistants, enquiry workflows, practical business automations and AI-tool integration concepts.
Keyword and content research, YouTube and publicly available social content research, campaign measurement, performance insights and digital analytics.
These are research and development directions. Scope, launch dates and commercial availability have not been announced.
Local LLM deployment, private document assistants, offline-first knowledge tools and device-aware AI implementation.
Agentic workflows, human-supervised AI operations, multi-step task automation, API integrations and workflow orchestration.
Local image and video generation, reproducible media pipelines, creative tooling and production workflow experimentation.
Browser and mobile applications, interactive learning experiences, games, prototyping and AI-enhanced development workflows.
Evidence-led SEO, AEO and GEO approaches, structured content, technical optimization and search visibility experiments.
Feasibility studies, technology evaluation and prototype exploration involving local models, automation and generative tools.
Optizord AI & Tech is an independent technology and creative innovation brand based in Kerala, India, with an international outlook. Its work spans browser games, interactive software, creative technology and hands-on AI experimentation.
Interactive projects: browser games and educational software concepts.
AI experiments: local models, generative workflows, agents and automation research.
Experience archive: website, creative and marketing services previously offered, with commercial availability currently paused.
Clear information about what is available today and what remains in development.
Throttle Outlaws and Circuit Rivals have browser-game links in the Current Highlights section. Android versions are planned; a Google Play listing link will be provided when a listing becomes available.
No. Regular commercial AI, web, creative and marketing services are currently paused. No date has been confirmed for their return.
Areas being explored include local language models, local AI video generation, agentic AI and intelligent automation. Earlier hands-on experiments include ComfyUI on an RTX 4060 system and local LLM text generation on older Intel CPUs.
Yes. Email optizordai@gmail.com with the topic and relevant details. An enquiry does not imply commercial availability or a commitment to take on work.
Have a question about a playable project, a technical experiment or a potential future collaboration? Choose the most relevant enquiry type. Standard commercial services remain paused.
Tell us what you are interested in. This form opens your email app with a prepared message; it does not upload your details or promise service availability.
Or email us directly at optizordai@gmail.com