Pasta-Devs/Marinara-Engine

AI frontend

What it solves

Marinara Engine provides a ready‑to‑run, locally hosted AI chat platform that lets users role‑play, run tabletop‑style games, and have casual conversations without dealing with complex setup or external accounts. It bundles model connections, visual assets, and optional AI agents so you can focus on storytelling and interaction.

How it works

The engine runs a local server (Docker, Windows, macOS/Linux, Android, or iOS PWA) and exposes a web UI. You pick a chat model (OpenAI, Anthropic, local Ollama, etc.) and optionally enable helper agents that track world state, generate images, or polish prose. The UI offers three modes—Conversation, Roleplay, and Game—each with customizable sprites, backgrounds, weather effects, and theme settings. All data stays on your device; authentication is optional for LAN or multiplayer sessions.

Who it’s for

  • Hobbyists and tabletop gamers who want AI‑driven NPCs and story generation.
  • Creators building character‑centric role‑play experiences without writing code.
  • Anyone looking for a plug‑and‑play local AI chat that can be extended with custom agents or themes.

Highlights

  • One‑click installation on desktop, Docker, Android, or as a PWA.
  • Three distinct chat modes with full‑body sprites, dynamic weather, and visual‑novel style displays.
  • Optional AI agents for world tracking, image illustration, music, and more; install only what you need.
  • Drag‑and‑drop prompt builder, lorebooks, and regex scripts for advanced prompt engineering.
  • Fully local storage (JSONL, Markdown, HTML export) and no required online account.
  • Extensive theming: fonts, colors, CSS themes, and movable UI windows.
  • Multiplayer support via explicit opt‑in and Basic Auth for private LAN sessions.
  • Wide provider support: OpenAI, Anthropic, Gemini, local Ollama, Stable Diffusion, ElevenLabs, etc.

Related

  • Project
    omnara-ai/omnara

    Omnira is an open‑source, self‑hostable platform for running AI agents on any LLM model and any machine. It provides a REST API, TypeScript SDK, and CLI to launch agents, stream events, and handle approvals. Features include model‑agnostic support, built‑in and custom tools, reusable skills, persistent memory, sub‑agents, multi‑tenant RBAC, scheduling, and integrations with Slack, Discord, GitHub, webhooks, and more. You can try it via Omnira Cloud or run it locally with Docker Compose; the Apache 2.0‑licensed code keeps all agent data in your own Postgres.

  • Project
    taylorsatula/mira

    Mira is an open‑source AI‑agent that maintains a single, never‑reset conversation thread, automatically summarises and stores older dialogue as first‑person memories, and uses a layered architecture (subcortical retrieval, Peanut Gallery metacognition, Text‑Based LoRA prompt adaptation) to stay context‑aware. It ships with a rich, on‑demand tool system (email, web search, Home Assistant, etc.), encrypted domain‑document handling, and a provider‑neutral LLM interface. Installation is a one‑line script that sets up a Python env, PostgreSQL, Vault, and optional model downloads, with update/backup commands for easy maintenance. Licensed under AGPL‑3.0.

  • Project
    intel/AI-Playground

    AI Playground is an open-source generative AI application suite for Intel AI PCs, offering offline chat, image generation, video generation, code assistance, and document search. It supports local models via Llama.cpp and OpenVINO, with an optional Hybrid Cloud mode for frontier models.

  • Project
    zhayujie/CowAgent

    CowAgent is an open‑source, self‑hosted AI assistant that plans tasks, runs tools, keeps a personal knowledge base, and can operate as a team of cooperating agents. It supports many LLM providers, multimodal input, dozens of chat channels (Telegram, Slack, WeChat, etc.), a plug‑in skill system, and a three‑tier memory with nightly “Deep Dream” distillation. Install via a one‑line script or Docker, manage everything through a web console, and extend the system through the public Skill Hub or custom tool integrations.

  • Project
    RunanywhereAI/runanywhere-sdks

    RunAnywhere SDKs provide a unified, on‑device AI runtime for phones, browsers, desktops and servers. A single C++ core routes calls to the best hardware engine (Snapdragon Hexagon NPU, Apple MLX, llama.cpp, ONNX, Core ML, CUDA, WebGPU, or CPU). Eight language‑specific SDKs (Swift, Kotlin, Flutter, React‑Native, Web, Electron, Python, and the `wally` CLI) expose the same high‑level API for LLM chat, vision‑language, speech‑to‑text, text‑to‑speech, embeddings, RAG, image generation, tool‑calling, and more, all offline and private.