Open source / For developers

Tools, experiments.
Code you can explore.

For developers and the technically curious. Projects built to solve a problem, test an idea, or understand something more deeply. Each has its own context, with the code available to explore.

AI & developer tools

Integrating AI, improving writing, controlling API traffic, and exploring backend development.

Go/LLMs/Multi-provider

Early development

LamarAI

One interface for AI integrations.

A type-safe Go SDK for connecting multiple LLM providers through a consistent interface. It brings generation, streaming, tool calling, and other model capabilities into an explicit integration layer.

  • Consistent provider interfaces and typed capabilities.
  • Error handling and implementation examples.
  • Early development; the API can change.

Rust/Swift/Flutter

macOS / Mobile in progress

ZaWriter

Writing assistance inside the active app.

An AI writing and voice assistant built around selected text and the application someone is already using. A shared Rust core handles the writing pipeline, while native interfaces handle platform integration.

  • Native Swift menu-bar app for macOS.
  • Rewrite and voice workflows with review before replacing text.
  • Flutter mobile interface in progress.

Python/Robyn/Redis

Upstream extension

Robyn Rate Limits Plus

More control over API traffic.

An extension of IdoKendo’s original Robyn Rate Limits project, adding strategies for controlling request volume and bursts in Python APIs. My contribution builds on that upstream work.

  • Fixed-window and token-bucket stores alongside sliding windows.
  • In-memory and Redis-backed options.
  • Configuration examples and algorithm tradeoffs.

Dart/APIs/Multi-isolate

Archived

Cruky

Exploring a backend experience for Dart.

A server-side framework for building APIs with annotation-based routes, middleware, and command-line tooling. The repository preserves an archived implementation.

  • JSON, form, and multipart request handling.
  • Multi-isolate execution, async routes, and development hot reload.
  • WebSockets, HTTPS, and static-file serving.

Applications & experiments

Engineering and learning through concrete problems, from URL matching and image classification to connected application workflows.

Dart/Routing/Data structures

Routing experiment

Routly

Data structures with a concrete job.

A tree-based routing experiment that matches URL paths against route templates. I used a practical backend problem to explore data structures and algorithms.

  • Dynamic path parameters such as /users/{id}.
  • Typed matching for integer parameters.
  • Published package with examples and tests.

Python/PyTorch/Computer vision

Computer-vision project

PetType AI

A focused image-classification experiment.

A computer-vision project that classifies pet types from images using a convolutional neural network. It applies deep-learning concepts to a specific image-recognition task.

  • Model implementation with Python and PyTorch.
  • Image handling with torchvision and Pillow.
  • Numerical processing with NumPy.

Flutter/FastAPI/MongoDB

Community application

Peti

A connected pet-community application.

A pet-focused social application joining a Flutter client, FastAPI backend, and MongoDB models. The repository brings the application, backend, and database diagrams together.

  • Registration, sign-in, and profile management.
  • Pet profiles, image and video posts, comments, and feeds.
  • API integration and database models with ODMantic.

Flutter/Dart/Operations

Management application

SpaceManager

Everyday space-management workflows.

A management application for coworking spaces and course centers. It connects rooms, reservations, courses, and students in one operational interface.

  • Rooms, sessions, and guest reservations.
  • Course and student workflows.
  • Administrative pages, a dashboard, and reusable packages.

Want to talk about a project?

Explore the README and code on GitHub, or reach me on WhatsApp.

Chat on WhatsApp ↗