OYYO is a proprietary AI workspace built by a team based in Dubai, designed to integrate native models, memory, tools, and agents into a single controlled AI work environment. Its core philosophy is "one intelligence, one workspace, one place to handle everything," letting users start from outcomes rather than repeating context, choosing models, and moving information between multiple tools.
Purpose and Positioning OYYO is aimed at scenarios that require completing a full work cycle. Users describe the outcome they want to achieve; OYYO Intelligence is responsible for understanding the goal, OYYO Prompt Knowledge improves instructions, OYYO Memory brings in relevant context, the runtime selects the permitted intelligence and execution mode, tools and agents carry out approved work, outputs are checked, provenance is retained, and useful knowledge is kept for next time. It is officially described as "models can only answer; OYYO is built to carry the work."
Main Capabilities and Layered Design OYYO is built in layers rather than as standalone applications. The layers publicly mentioned include:
- OYYO Work: real deliverables such as documents, presentations, spreadsheets, reports, and plans (roadmap 0.8).
- OYYO Growth: research, brand, strategy, campaigns, and localization, carried out in persistent context.
- OYYO Research: current information, source comparison, and retained citations (in development, 0.4 Knowledge).
- OYYO Code: understanding codebases, planning changes, writing, reviewing, testing, and documentation (roadmap 0.9 Agents).
- OYYO Creative: visual direction, concepts, storyboards, and multimodal creative work (roadmap 0.6-0.7 Vision and Media).
- OYYO Memory: persistently accumulated knowledge, source-backed facts, corrections, terminology, and Prompt Knowledge, reducing repeated explanations (in development, 0.4 Knowledge).
- OYYO Language: semantic translation with correctable translation memory (roadmap 0.5 Language).
- OYYO Agents: tools, code, and browser, placed behind permissions, sandboxes, and audit trails (roadmap 0.9 Agents).
- OYYO Runtime: CPU-first local execution, portable workspace, and hardware-aware acceleration direction (in development, 0.2-0.3 Runtime and Adaptive).
EnablingHow to use Users describe goals in natural language, for example "launch our new smartwatch in France." The system breaks it down into steps such as researching the market and competitors, identifying the audience, defining positioning, building the campaign strategy, creating a channel plan, writing campaign copy, translating and localizing into French, generating visual concepts, creating a social calendar, building a launch presentation, creating a budget spreadsheet, and saving campaign knowledge for future use. Sensitive steps stay local, while heavy steps are sent to the cloud to balance speed, privacy, and cost. Each deliverable retains its provenance, decisions, and the run record that generated it. Users retain control over execution mode; OYYO makes suggestions, and the user decides.
Execution modes and privacy OYYO is designed to choose among local, private, hybrid, and approved cloud execution based on the task, hardware, privacy rules, latency, quality, and cost. The portable workspace direction makes knowledge, files, deliverables, and eligible local model assets movable without forcing cloud sync. Examples include: private document analysis using a local runtime, with data staying on the device; long-text reasoning using a cloud model for maximum capability; and batch classification using a small local model, which is fast and low-cost.
Memory and corrections OYYO Memory is designed to retain approved facts, entities, relationships, decisions, files, terminology, corrections, procedures, and results, with scope and provenance. If a user corrects something once, the correction is retained and reused. For example, customer entities, an approved glossary, automatically applied corrections, pricing decisions from board materials, and so on.
Agents and boundaries OYYO Agents use tools, code, browsers, connectors, and MCP through explicit capability boundaries. Permission to read information and permission to change information are different, and significant operations may require approval. For example, drafting a supplier email is "proposed," updating a campaign spreadsheet is "approved," and sending to 4,200 contacts is "requires approval." It is officially emphasized that without a human decision, nothing irreversible happens.
Native models and benchmarks OYYO is developing a proprietary family of native AI models designed specifically for the OYYO runtime, memory, tools, agents, language intelligence, and real business workflows. The goal is to be independent but not isolated: the native models are intended to support the declared OYYO capabilities without permanently depending on external model providers; when other qualified local or cloud models can improve a task and policy allows it, OYYO Intelligence can also use that capability. OYYO Benchmark is being built to measure real task success rates, quality, repeatability, and hardware cost, covering general reasoning, business execution, coding and terminal work, tools/functions/agents, documents/spreadsheets/presentations, language and semantic translation, memory and temporal knowledge, image/audio/video understanding, growth and marketing execution, security and reliability, hardware efficiency, and other capability families.
Engineering status and limitations According to official information, OYYO Core Runtime 0.2.0 dev.1 is in active private development and has been used internally for real development and testing, including local inference, model lifecycle fundamentals, API compatibility, streaming, embeddings, connector fundamentals, controlled tool execution, hardware reporting, and provenance fundamentals. Native models, benchmarks, active knowledge, Prompt Knowledge, Operator Studio, portability, and a broader product experience are still being advanced in parallel. OYYO is not yet generally available, public release is not yet available, and all roadmap capabilities are explicitly labeled. OYYO ONE AI was showcased by HRN INNOVATION TECHNOLOGIES LTD at AI Everything Abu Dhabi 2026.
Use cases OYYO is aimed at business work that requires research, creation, tool use, memory, and controlled execution within a single persistent context, such as market research, brand strategy, campaign planning, localization, document and presentation creation, code understanding and changes, visual concept generation, and more. Its design emphasizes business-native, local-first, evidence before brand, and learning from corrections.
Limitations and considerations OYYO is currently in private development, has not yet been publicly released, and many capabilities are still on the roadmap. Official specific pricing, platform support details, or public release dates have not been provided. Users should follow its early access list and roadmap for the latest information.
