OYYO investor access
Do not fund another chatbot. Fund the intelligence layer that can make AI operational.
OYYO is a proprietary AI project conceived and architected in Dubai by Vincenzo Picciuolo and developed with his team through HRN INNOVATION TECHNOLOGIES LTD, DIFC. OYYO is building an intelligence and execution platform that can run its own native models, understand organizational context, improve task instructions, use tools, create real work, operate under permissions and adapt across local, private, hybrid and cloud infrastructure.
Current engineering baseline, OYYO 0.2 Runtime, active private development and daily internal use.
US$3.5M proposed seed round
Last updated,
In ten seconds
What if the AI was not another app your company had to operate?
Today people operate AI. They choose models, engineer prompts, copy context, move data between apps, assemble deliverables and connect the systems themselves.
OYYO changes the boundary. The operator describes the outcome.
OYYO Intelligence is designed to prepare the instruction, bring in the right knowledge, select permitted intelligence and tools, execute the work, verify the result and preserve what should be remembered.
The human should define the outcome. The intelligence layer should manage the complexity.
Engineering status
The round accelerates the project. It does not start it.
OYYO already has a working 0.2 Runtime development line. The current system is running internally and is used every day by the founder and development team.
It already contains real local model execution, local model lifecycle management, an all in one CLI, OpenAI compatible API foundations, streaming responses, embeddings, connector foundations, controlled write approvals, hardware capability reporting and provenance foundations. The next phase turns that engineering base into the complete OYYO product experience. OYYO 0.3 Adaptive is the next milestone, not completed work.
Local model lifecycle
Verified model activation, cryptographic integrity and local registry foundations.
Real local inference
CPU first execution today, optional acceleration paths and adaptive hardware work next.
All in one runtime
One oyyo executable is being built as the operator path for model management, inference, connectors and the local API.
OpenAI compatible interface
Compatibility surfaces for models, chat, responses and embeddings reduce developer migration friction.
Connectors and tools
Built in catalog foundations cover major productivity, communication, storage, development, CRM, project, database and protocol classes. Broader provider adapters remain active development work.
Controlled actions
Read access can be automated under policy. Consequential writes can require explicit approval tied to the exact action.
Provenance
Generated outputs can carry non visible OYYO provenance metadata with detached generation receipts.
OYYO Core Runtime
0.2.0 runtime dev.1, active private development
OYYO Native Models
Foundation programme active, model family and qualification architecture defined
OYYO Benchmark
Foundation framework active
OYYO SDK
Public foundation interfaces for Python and TypeScript
Operator Studio
Product experience in development
Public release
Not yet public
Parallel programmes
Three systems are now moving forward together.
OYYO Core Runtime
The execution system, hardware abstraction, model lifecycle, local API, tools, connectors, approvals, provenance and orchestration foundation.
OYYO Native Models
The proprietary native intelligence family designed to make OYYO independent from any single external AI provider.
OYYO Benchmark
The qualification system that measures models and runtimes on real work, quality, tools, language, artifacts, security, reliability and hardware efficiency before promotion.
Build the intelligence. Build the runtime. Measure both.
Independence thesis
Disconnect every external AI provider. OYYO should still work.
This is one of the defining engineering principles behind OYYO Native Models. External models can be valuable collaborators, but they should not be a permanent dependency for the base OYYO product.
An OYYO release that claims independent native intelligence must be able to perform its declared core workflows using OYYO native models with external model providers disabled.
This is not anti cloud. It is anti dependency.
Compounding
Every improvement in AI can become an OYYO improvement.
OYYO is designed so models are powerful engines inside a larger operating system for intelligence.
When OYYO native models improve, the platform improves. When a stronger local model becomes available and is approved, OYYO Intelligence can use it. When a specialist cloud model is useful and policy allows it, OYYO can route selected work to it.
The runtime remains responsible for context, memory, policy, permissions, tools, workflow state and the final outcome. Model progress does not have to make OYYO obsolete. It can make OYYO stronger.
The work loop
One request can become an entire verified work loop.
01
Understand
Interpret the goal and constraints.
02
Context
Bring in the authorized files, knowledge, conversations, systems and memory.
03
Select
Choose the appropriate OYYO model, specialist, local model or approved cloud capability.
04
Plan
Build the work path.
05
Execute
Use tools, agents, connectors and artifact engines.
06
Verify
Check outputs, evidence, structure and execution state.
07
Remember
Preserve approved knowledge, corrections, decisions and provenance for future work.
Native models
OYYO is not building native models for a bigger number on a leaderboard.
The OYYO model programme is designed around useful capability on real hardware. The optimization target is not parameter count alone.
OYYO evaluates capability per GB of memory, per watt, per unit of latency, per serving cost and per real task completed. That matters because OYYO is designed to run in more than one place, laptop, workstation, private server, enterprise infrastructure, cloud and hybrid environment. The runtime should select the strongest qualified configuration that fits the device, task, context, policy and performance requirement.
Memory
The longer OYYO works with an organization, the less that organization should need to explain.
OYYO Memory is designed as cumulative organizational knowledge, not just a scrollback of chats.
It can preserve facts, entities, products, customers, projects, relationships, decisions, assets, corrections, preferred terminology, outcomes, versions, conflicts, permissions and provenance. A model response disappears. Organizational knowledge compounds. This is product direction, not a claim that every stage is already released.
Day 1
Understands the task.
Month 1
Understands how the organization speaks and works.
Quarter 1
Connects patterns across projects, decisions and outcomes.
Year 1
Becomes a deeply contextual intelligence layer for the organization.
Control
Useful automation does not require uncontrolled automation.
OYYO separates intelligence from authorization. A model can propose an action without being allowed to authorize itself.
Safe reads can run under policy. Consequential writes can require explicit approval tied to the exact action and arguments. Tool outputs are treated as external data until handled by the orchestration and policy layer. The target is practical automation with permissions, auditability and human control where consequences matter.
Provenance
Intelligence should leave evidence.
OYYO is building provenance into the runtime rather than adding it as an afterthought.
The 0.2 development line includes non visible provenance metadata and detached hash based generation receipts. The long term direction includes verifiable provenance across generated outputs and exported business artifacts.
OYYO Invisible Provenance Watermark
Designed to carry OYYO generation identity through non visible channels appropriate to the output format.
Deployment
Local when it matters. Cloud when it helps.
OYYO is designed so the workload, policy and organization determine where intelligence runs.
Local
Sensitive knowledge, private work, routine inference, offline use and owned hardware.
Private
Company controlled infrastructure and restricted environments.
Hybrid
Keep selected context and policy local while using approved hosted capability for selected tasks.
Cloud
Managed capacity, large reasoning workloads, specialist intelligence and media workloads when selected.
Choice becomes part of the architecture.
Defensibility
The moat is not one model checkpoint.
Native intelligence
A proprietary OYYO model family that can power the platform independently.
Runtime distribution
A portable execution and workspace direction for knowledge, files, artifacts, settings and eligible model assets across supported clients without mandatory cloud synchronization.
Hardware adaptation
Model and configuration selection based on actual device capability rather than one required infrastructure path.
Organizational memory
Cumulative knowledge, Prompt Knowledge, reusable procedures, corrections, provenance and relationships.
Connectors and actions
Deep integration, permissions, approvals and execution semantics.
Business workflows
Reusable operating patterns across departments and outcomes.
Artifacts and media
Documents, slides, spreadsheets, code, image, voice, audio and video as real work products.
Benchmark discipline
Qualification based on measured performance and operational suitability rather than vendor marketing.
Trust layer
Provenance, auditability, deployment choice and controlled action.
Defensibility compounds as runtime depth, native intelligence, integrations, organizational context and operator trust accumulate together.
Roadmap
Build the foundation correctly, then compound capability.
Milestones are published without invented completion dates.
0.1 Foundation
Baseline complete.
0.2 Runtime
Active private development, current engineering line.
Native Models programme
Now active in parallel with Runtime and Benchmark.
0.3 Adaptive
Real accelerator probes, offload planning and AutoTune.
Knowledge
Durable memory, entity and fact structures, permissions and provenance.
Language
Semantic and contextual language intelligence with translation memory.
Vision and Media
Image, document, audio and video understanding and production workflows.
Work
Validated business artifacts such as documents, slides, spreadsheets and PDF.
Agents
Production grade controlled action and workflow experiences.
1.0 Stable
Integrated stable OYYO product objective.
Seed round
US$3.5M to accelerate OYYO from Runtime toward Stable.
The objective of the round is not to start an AI experiment. The objective is to accelerate a working engineering foundation into an integrated product that organizations can deploy, use and pay for.
45 percent
Engineering and AI runtime
18 percent
Product and Operator Studio
12 percent
Security, compliance and IP
10 percent
Infrastructure and benchmarks
10 percent
Go to market and pilots
5 percent
Operations and contingency
- US$3.5M proposed seed financing target
- 18 month disciplined execution runway target
- OYYO 1.0 integrated stable product objective
Use of capital
Capital turns velocity into separation.
01
Harden the 0.2 Runtime
Packaging, release matrix, SDK depth, provider adapters, security and production hardening.
02
Deliver OYYO Adaptive
Real hardware probing, offload planning, AutoTune and broader acceleration support.
03
Accelerate OYYO Native Models
Native model architecture, training and adaptation pipeline, qualification, packaging, quantization and deployment integration.
04
Build Operator Studio
The simple Ask, Build and Deploy experience for operators who should not need to become AI engineers.
05
Make knowledge durable
Memory, entities, provenance, permissions, translation memory and organizational context.
06
Enterprise pilots
Private deployments, department intelligence, controlled agents, integrations and repeatable commercial playbooks.
Origin
Built in Dubai. Made in UAE.
OYYO is not a foreign AI product being relabeled for the region.
The project was conceived, architected and developed in Dubai by Vincenzo Picciuolo, his team and HRN INNOVATION TECHNOLOGIES LTD in DIFC. The ambition is global. The engineering origin is Dubai.
Public engineering proof
The public surface is small by design. The technology is proprietary by design.
OYYO is proprietary and not open source. Selected public repositories expose model metadata, benchmark methodology and SDK integration surfaces so developers, partners and investors can inspect the engineering philosophy without publishing the private OYYO Core Runtime or proprietary model development internals.
OYYO Models
Public model registry, family targets, compatibility metadata, provenance and release qualification records.
https://github.com/vpicciuolo/oyyo-models
OYYO Benchmark
Public benchmark methodology for reasoning, business execution, coding, tools, agents, language, memory, security and hardware efficiency.
https://github.com/vpicciuolo/oyyo-benchmark
OYYO SDK
Public Python and TypeScript integration surfaces for OYYO compatible and OYYO native APIs.
https://github.com/vpicciuolo/oyyo-sdk
Founder GitHub
Public GitHub profile of Vincenzo Picciuolo.
https://github.com/vpicciuolo
Engineering update
Engineering update, 15 September 2026
The downloadable Investor Product Dossier captures the OYYO engineering baseline as of 9 September 2026. Since that baseline, the OYYO Native Models programme has formally moved into active parallel development with OYYO Core Runtime and OYYO Benchmark.
This website reflects the latest public positioning and current engineering direction.
If this is within your investment thesis, speak with us directly.
OYYO is opening conversations with investors who understand infrastructure, applied AI, enterprise software and the opportunity to build globally relevant AI technology from Dubai. Download the dossier for the detailed architecture, roadmap and proposed seed round, or request a direct investor conversation with the team.
Investor product dossier
OYYO investor presentation 2026
PDF · 1.5 MB · Baseline 9 September 2026
Last updated,
Investor questions
Answers before the conversation.
What is OYYO?
OYYO is a proprietary AI project and platform developed in Dubai by HRN INNOVATION TECHNOLOGIES LTD. It combines native OYYO models, a portable AI runtime, persistent knowledge, tools, agents, multimodal capabilities and controlled business execution.
Is OYYO another chatbot?
No. Chat is one possible interface. OYYO is being built as an intelligence and execution platform that can understand objectives, use company context, select models, work through tools and deliver finished outcomes.
Does OYYO have its own AI models?
Yes. OYYO is developing its own proprietary native AI model family. The family is designed to work directly with the OYYO runtime and to support local, private and cloud execution where hardware permits.
Does OYYO depend on OpenAI, Anthropic, Google or another model provider?
No single external provider is intended to be a permanent dependency for the base OYYO product. OYYO can collaborate with external models when useful and permitted, but the native model programme exists so OYYO can operate independently for its declared capabilities.
Is OYYO open source?
No. OYYO is proprietary technology. Selected public repositories expose benchmark methodology, model metadata, SDK interfaces and compatibility information, but the core runtime, orchestration engine and proprietary model development internals remain private.
Can OYYO run locally?
Yes, local execution is a core architectural direction. The current 0.2 Runtime already supports real local inference through a CPU first path. Broader adaptive accelerator selection is part of the next engineering milestone.
What stage is OYYO at today?
OYYO is in active private development. The 0.2 Runtime line is already running internally and used in daily team testing. The Native Models, Benchmark and Runtime programmes are moving forward in parallel. OYYO is not publicly released yet.
Where is OYYO being built?
OYYO was conceived, architected and is being developed in Dubai, UAE, by Vincenzo Picciuolo, his team and HRN INNOVATION TECHNOLOGIES LTD in DIFC.
What is the current investment round?
OYYO is presenting a proposed US$3.5M seed financing target to accelerate the current runtime foundation toward the integrated OYYO 1.0 product objective and enterprise deployment.