Cyréna is an AI-native desktop engineering workspace that adapts to your workflow through extensions.
It runs alongside your IDE, not inside it. Extensions let you install only what you need instead of shipping a massive monolithic app.
Cyréna helps developers build, maintain, and evolve real software projects across multiple engineering domains including .NET, embedded systems, firmware, web platforms, and static websites.
You choose the model. Cyréna orchestrates the workflow.
Download (alpha) | Cyréna HUD (Windows) | Website | Docs | Screenshots | My Model Experience
Cyréna was built from a practical problem.
AI app builders could generate software quickly, but the infrastructure, hosting, and execution environment remained locked behind proprietary platforms. The software was not truly mine.
Cyréna solves that by helping developers move faster with AI while keeping full control over their code, architecture, infrastructure, and deployment.
The goal is not magic.
The goal is disciplined AI-assisted engineering inside real projects.
Cyréna is a desktop application for AI-assisted engineering.
It works with OpenAI and Ollama, allowing you to use cloud or local models and switch between them depending on the task.
Cyréna is designed to slot into your existing workflows and tools, enhancing and accelerating them rather than disrupting them.
It can:
- implement features
- modify existing codebases
- repair build failures
- generate documentation
- retain project-specific knowledge
- support repeatable engineering workflows
Cyréna is built around extensions.
Extensions allow support for different engineering domains such as .NET, Angular, Arduino, PlatformIO, and static websites.
The extension model also allows Cyréna to expand beyond software engineering into entirely new domains over time.
Developers remain in control of their code, infrastructure, and workflows.
👉 .Net(YouTube) 👉 Arduino(YouTube)
This is not autocomplete.
This is structured AI-assisted engineering inside real projects.
Cyréna is built around extensions.
Every engineering domain is implemented as an extension, including:
- .NET
- Arduino IDE
- PlatformIO
- Angular
- Static Website Development
Extensions define:
- project structures
- prompts
- workflows
- tooling constraints
- repair strategies
- domain-specific behavior
This allows Cyréna to adapt to different workflows without becoming a generic “AI assistant.”
Cyréna dynamically updates system prompts based on:
- active extensions
- project type
- enabled capabilities
- engineering domain
- current workflow context
The model only receives the instructions relevant to the active task.
This keeps prompts focused, smaller, and domain-aware.
Features can be enabled or disabled per chat.
Disabled functionality is removed from the model’s available toolset entirely.
This prevents:
- irrelevant tool usage
- accidental actions
- prompt bloat
- workflow confusion
The AI only knows what it needs to know.
Cyréna supports prompt queuing.
Developers can queue multiple engineering tasks and allow the platform to work through them sequentially.
Each iteration:
- Loads relevant context
- Inspects project files
- Performs the requested work
- Builds and validates
- Repairs issues if necessary
- Persists technical knowledge
- Continues to the next queued task
This allows longer engineering workflows to execute in a controlled and repeatable way.
Cyréna currently supports multiple engineering domains through extensions.
- C# Class Libraries
- Blazor Applications
- MVC Applications
- MVC Libraries
-
Arduino IDE
-
PlatformIO
- Arduino Framework
- ESP-IDF
- Angular
- Static HTML/CSS/JavaScript websites
Each domain has its own:
- project structures
- prompts
- engineering rules
- architecture constraints
- repair workflows
Different domains. Same engineering discipline.
API References are structured technical documents generated and maintained by the platform.
They provide persistent technical memory describing:
- APIs
- architecture rules
- service contracts
- module behavior
- integration patterns
- implementation details
API References are grounded in real project code rather than hallucinated summaries.
This allows Cyréna to maintain long-term project understanding beyond normal chat context limitations.
Sticky Notes act as lightweight persistent project memory.
They capture:
- architectural decisions
- engineering constraints
- developer reminders
- project-specific rules
- workflow guidance
Sticky Notes survive chat resets and model switches.
Cyréna works directly inside real project structures.
It:
- reads existing files
- inspects architecture
- follows project conventions
- performs minimal edits
- validates builds
- repairs failures iteratively
The platform is designed to avoid uncontrolled code generation and reduce architectural entropy over time.
Engineering tasks operate through structured loops:
- Inspect project state
- Load API References
- Read Sticky Notes
- Review relevant files
- Implement minimal changes
- Build and validate
- Repair failures
- Persist technical knowledge
- Summarise work completed
This approach keeps workflows controlled and repeatable.
Cyréna is designed to work with:
- Ollama
- OpenAI
- compatible AI providers
Developers can:
- run models locally
- switch providers mid-project
- work offline
- keep code on their own infrastructure
The platform is model agnostic.
You bring the model. Cyréna provides the engineering workflow.
Cyréna now includes a static website engineering extension.
The Website extension supports:
- semantic HTML5
- responsive CSS
- vanilla JavaScript
- structured project layouts
- SEO-friendly static content
- accessibility-aware markup
- asset management
- multi-page websites
This extension was used to rebuild the Cyréna marketing website from Angular-rendered landing pages into static crawlable content.
- .NET 10
- Windows / Linux / macOS
- Ollama or OpenAI
- Recommended: ≥16k context models
Hardware requirements scale with model size.
- Run Cyréna
- Open Settings
- Add an Ollama connection
- Configure model + tokens + context
- Save
- Open Settings
- Add API key + model
- Save
- Click New Chat
- Select a project type
- Configure the project
- Submit
Baseline development system:
- RTX 3060 12GB
- 48GB RAM
- Ryzen 7 8700F
Typical local test configuration:
gpt-oss:20b- 8k tokens
- 16k context
Larger models require additional VRAM and RAM.
Cyréna is being built around:
- disciplined AI behavior
- reversible engineering workflows
- project-aware reasoning
- persistent technical memory
- predictable code generation
- developer supervision
- architecture safety
- long-term maintainability
The goal is not replacing developers.
The goal is building better engineering workflows.
Cyréna stands on the shoulders of excellent open-source projects:
Cyréna modifies real files inside real projects.
Always use version control. Always review AI-generated changes.
You are responsible for your codebase.
Provided as-is. Use at your own risk.