[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"project-80725":3},{"id":4,"name":5,"fullName":6,"owner":7,"repo":5,"description":8,"homepage":9,"htmlUrl":10,"language":11,"languages":10,"totalLinesOfCode":10,"stars":12,"forks":13,"watchers":14,"openIssues":15,"contributorsCount":16,"subscribersCount":16,"size":16,"stars1d":16,"stars7d":17,"stars30d":18,"stars90d":16,"forks30d":16,"starsTrendScore":16,"compositeScore":19,"rankGlobal":10,"rankLanguage":10,"license":20,"archived":21,"fork":21,"defaultBranch":22,"hasWiki":23,"hasPages":23,"topics":24,"createdAt":10,"pushedAt":10,"updatedAt":44,"readmeContent":45,"aiSummary":46,"trendingCount":16,"starSnapshotCount":16,"syncStatus":47,"lastSyncTime":48,"discoverSource":49},80725,"Open-Mythos-2","creatorofsomethingthatisgood\u002FOpen-Mythos-2","creatorofsomethingthatisgood","This is an attempt to make original open mythos better with chat ui.","https:\u002F\u002Fcreatorofsomethingthatisgood.github.io\u002FOpen-Mythos-2\u002F",null,"Python",67,4,18,1,0,5,26,2.1,"The Unlicense",false,"main",true,[25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43],"21","agent","agentic-ai","ai","best","fast","future","growing","home","hot","linux","mac","madebykid","mythos","open-mythos","open-mythos-2","popular","python","stared","2026-06-12 02:04:06","\u003Cdiv align=\"center\">\n\n\u003Cpre>\n                  ██████╗ ██████╗ ███████╗ ███╗   ██╗\n                 ██╔═══██╗██╔══██╗██╔════╝ ████╗  ██║\n                 ██║   ██║██████╔╝█████╗   ██╔██╗ ██║\n                 ██║   ██║██╔═══╝ ██╔══╝   ██║╚██╗██║\n                 ╚██████╔╝██║     ███████╗ ██║ ╚████║\n                  ╚═════╝ ╚═╝     ╚══════╝╚═╝  ╚═══╝\n███╗   ███╗██╗   ██╗████████╗██╗  ██╗ ██████╗ ███████╗       ██████╗\n████╗ ████║╚██╗ ██╔╝╚══██╔══╝██║  ██║██╔═══██╗██╔════╝       ╚════██╗\n██╔████╔██║ ╚████╔╝    ██║   ███████║██║   ██║███████╗  █████╗ █████╔╝\n██║╚██╔╝██║  ╚██╔╝     ██║   ██╔══██║██║   ██║╚════██║  ╚════╝██╔═══╝\n██║ ╚═╝ ██║   ██║      ██║   ██║  ██║╚██████╔╝███████║       ███████╗\n╚═╝     ╚═╝   ╚═╝      ╚═╝   ╚═╝  ╚═╝ ╚═════╝ ╚══════╝       ╚══════╝\n\u003C\u002Fpre>\n\n> *\"From the void of computation, a new oracle speaks — no gods, no cloud, only the terminal.\"*\n\n\u003Cbr>\n\n\u003Cimg src=\"https:\u002F\u002Freadme-typing-svg.demolab.com?font=Fira+Code&weight=700&size=22&pause=1000&color=FFD700&center=true&vCenter=true&width=700&lines=The+Oracle+runs+in+your+terminal.;No+API.+No+cloud.+No+gods.;Pure+local+AI+power.;OpenMythos-2+has+awakened.\" alt=\"Typing SVG\" \u002F>\n\n\u003Cbr>\n\n\u003Cimg width=\"845\" height=\"372\" alt=\"OpenMythos-2 in action\" src=\"https:\u002F\u002Fgithub.com\u002Fuser-attachments\u002Fassets\u002F8e3b6a5b-117a-41c6-afbb-394e5cb8c023\" \u002F>\n\n\u003Cbr>\n\n[![npm published](https:\u002F\u002Fimg.shields.io\u002Fnpm\u002Fv\u002Fopen-mythos-2?style=for-the-badge&label=npm&color=cb3837)](https:\u002F\u002Fwww.npmjs.com\u002Fpackage\u002Fopen-mythos-2)\n[![GitHub stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fcreatorofsomethingthatisgood\u002FOpen-Mythos-2?style=for-the-badge&logo=github&color=FFD700&label=Stars)](https:\u002F\u002Fgithub.com\u002Fcreatorofsomethingthatisgood\u002FOpen-Mythos-2\u002Fstargazers)\n[![GitHub forks](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fforks\u002Fcreatorofsomethingthatisgood\u002FOpen-Mythos-2?style=for-the-badge&logo=github&color=9cf&label=Forks)](https:\u002F\u002Fgithub.com\u002Fcreatorofsomethingthatisgood\u002FOpen-Mythos-2\u002Fnetwork\u002Fmembers)\n[![License: MIT](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FLicense-MIT-gold?style=for-the-badge)](LICENSE)\n[![No API Required](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FNo%20API-Required-darkgreen?style=for-the-badge&logo=gnubash&logoColor=white)]()\n[![Runs in Terminal](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FRuns%20In-Terminal-black?style=for-the-badge&logo=windowsterminal&logoColor=white)]()\n[![Open Source](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FOpen-Source-orange?style=for-the-badge&logo=opensourceinitiative&logoColor=white)]()\n[![Python](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FPython-3.10+-3776AB?style=for-the-badge&logo=python&logoColor=white)]()\n\n\u003Cbr>\n\n**Fully local, offline AI in your terminal. Zero API keys. Zero cloud. Zero limits.**\n\n[Get started in 30 seconds](#-getting-started) · [Features](#-features) · [Commands](#-command-reference) · [RML](#-rml--reinforcement-machine-learning) · [Contributing](#-contributing)\n\n\u003C\u002Fdiv>\n\n---\n\n## Why OpenMythos-2?\n\n| | Cloud AI (ChatGPT, Claude, etc.) | **OpenMythos-2** |\n|---|---|---|\n| **Privacy** | Your data leaves your machine | **Everything stays local** |\n| **Cost** | Monthly subscription or per-token fees | **Free forever** |\n| **Offline** | Requires internet | **Works anywhere** |\n| **Setup** | API keys, accounts, billing | **Just run it** |\n| **Customization** | Locked persona & behavior | **RML adapts to you** |\n| **Security scanning** | Separate tool needed | **Built-in SAST + AI audit** |\n\n---\n\n## ⚡ What is OpenMythos-2?\n\n**OpenMythos-2** is a fully **local**, **offline** AI that lives and breathes inside your terminal. Inspired by the ancient myths of oracles, gods, and forgotten wisdom — it brings the power of intelligent conversation and reasoning directly to your command line, with **zero API keys**, **zero cloud dependency**, and **zero limits**.\n\nLike Prometheus stealing fire from the gods, OpenMythos-2 brings the fire of AI to every machine.\n\n---\n\n## 🏛️ Features\n\n| Feature |  Description |\n|---|---|\n|  **No API Required** | Fully local — nothing ever leaves your machine |\n|  **Terminal Native** | Built from the ground up for the command line |\n|  **Intelligent Reasoning** | Context-aware, multi-turn conversations |\n|  **Offline First** | Works anywhere — planes, bunkers, the underworld |\n|  **Mythos Persona** | Answers in the voice of an ancient, wise oracle |\n|  **Lightweight** | Minimal dependencies, blazing fast startup |\n|  **Open Source** | Fully transparent and community-driven |\n|  **Private by Design** | Your conversations are yours alone |\n|  **RML (Reinforcement ML)** | Learns from your feedback and adapts — edits system prompt and generation params to match your preferences | when command is \u002Frml on |\n\n##  The Myth Behind the Machine\n\nThe name draws from **mythos** (μῦθος) — the ancient Greek word for *story*, *legend*, and *the spoken word of truth*. OpenMythos-2 embodies that spirit: a storyteller, reasoner, and companion that runs **entirely on your machine**, with no external gods (servers) to pray to.\n\n---\n\n##  Getting Started\n\n### Prerequisites\n\n- Python `3.10+`\n- `pip`\n- A terminal (bash, zsh, PowerShell, cmd)\n\n###  Installation\n\nPick your platform and package manager:\n\n---\n\n\u003Cdetails>\n\u003Csummary>\u003Cstrong> Linux — Git Clone (recommended)\u003C\u002Fstrong>\u003C\u002Fsummary>\n\n**Ubuntu \u002F Debian:**\n\n```bash\n# Clone the sacred repository\ngit clone https:\u002F\u002Fgithub.com\u002Fcreatorofsomethingthatisgood\u002FOpen-Mythos-2.git\ncd Open-Mythos-2\n\n# One-time setup (installs deps, builds llama-cpp-python with Vulkan\u002FCPU fallback)\nsudo bash setup.sh\n\n# Download the model (~4.5 GB, first time only)\nmythos model download\n\n# Start chatting\nmythos\n```\n\n> **Note:** `.\u002Fsetup.sh` is optimized for Fedora (`dnf`). On Ubuntu\u002FDebian, it will warn but still work — system deps may need manual install:\n> ```bash\n> sudo apt install -y python3 python3-pip python3-dev gcc g++ make cmake git libopenblas-dev libvulkan-dev mesa-vulkan-drivers\n> ```\n\n**Arch Linux:**\n\n```bash\ngit clone https:\u002F\u002Fgithub.com\u002Fcreatorofsomethingthatisgood\u002FOpen-Mythos-2.git\ncd Open-Mythos-2\n\n# Install system deps first\nsudo pacman -S --needed python python-pip gcc make cmake vulkan-headers vulkan-icd-loader openblas\n\n# Run setup\nsudo bash setup.sh\nmythos model download\nmythos\n```\n\n**Fedora:**\n\n```bash\ngit clone https:\u002F\u002Fgithub.com\u002Fcreatorofsomethingthatisgood\u002FOpen-Mythos-2.git\ncd Open-Mythos-2\n\n# One-command setup (Fedora-optimized, includes Vulkan GPU support)\nsudo bash setup.sh\nmythos model download\nmythos\n```\n\n**Offline \u002F Air-gapped machines:**\n\n```bash\n# On a connected machine:\n.\u002Fscripts\u002Fmythos-export-data.sh # creates offline-bundle\u002F (~4.5 GB)\n\n# Transfer the bundle to the target machine, then:\n.\u002Fscripts\u002Fmythos-import-data.sh .\u002Foffline-bundle\n```\n\n\u003C\u002Fdetails>\n\n\u003Cdetails>\n\u003Csummary>\u003Cstrong> macOS — Git Clone (recommended)\u003C\u002Fstrong>\u003C\u002Fsummary>\n\n```bash\n# Clone the sacred repository\ngit clone https:\u002F\u002Fgithub.com\u002Fcreatorofsomethingthatisgood\u002FOpen-Mythos-2.git\ncd Open-Mythos-2\n\n# One-time setup (Metal GPU on Apple Silicon, CPU on Intel Macs)\n.\u002Fsetup-macos.sh\n\n# Download the model (~4.5 GB, first time only)\n.\u002Fmythos model download\n\n# Start chatting\n.\u002Fmythos\n```\n\n> **Note:** On Apple Silicon (M1\u002FM2\u002FM3\u002FM4), the setup builds llama-cpp-python with **Metal GPU acceleration** for fast inference. Intel Macs fall back to CPU mode. [Homebrew](https:\u002F\u002Fbrew.sh\u002F) is recommended for build dependencies (`brew install cmake`).\n\n**via Homebrew (tap):**\n\n```bash\nbrew tap creatorofsomethingthatisgood\u002Ftap\nbrew install open-mythos-2\n\nmythos model download\nmythos\n```\n\n\u003C\u002Fdetails>\n\n\u003Cdetails>\n\u003Csummary>\u003Cstrong> Windows — Git Clone (recommended)\u003C\u002Fstrong>\u003C\u002Fsummary>\n\n```powershell\n# Clone the sacred repository\ngit clone https:\u002F\u002Fgithub.com\u002Fcreatorofsomethingthatisgood\u002FOpen-Mythos-2.git\ncd Open-Mythos-2\n\n# One-time setup (PowerShell — detects CUDA\u002FVulkan\u002FCPU automatically)\n.\\setup-windows.ps1\n\n# Download the model (~4.5 GB, first time only)\n.\\mythos.bat model download\n\n# Start chatting\n.\\mythos.bat\n```\n\n**Prerequisites:**\n- **Python 3.10+** — [python.org](https:\u002F\u002Fwww.python.org\u002Fdownloads\u002F) (check \"Add Python to PATH\" during install)\n- **Microsoft C++ Build Tools** — [visualstudio.microsoft.com](https:\u002F\u002Fvisualstudio.microsoft.com\u002Fvisual-cpp-build-tools\u002F) (select \"Desktop development with C++\" workload)\n- **CMake** — `winget install Kitware.CMake` or [cmake.org](https:\u002F\u002Fcmake.org\u002Fdownload\u002F)\n- **GPU acceleration (optional):**\n  - **NVIDIA:** Install [CUDA Toolkit](https:\u002F\u002Fdeveloper.nvidia.com\u002Fcuda-downloads) for CUDA support\n  - **AMD\u002FIntel:** Install [Vulkan SDK](https:\u002F\u002Fvulkan.lunarg.com\u002Fsdk\u002Fhome) for Vulkan support\n\n> **Note:** If you skip the build tools, the setup will attempt a prebuilt wheel (CPU-only). For GPU acceleration, the C++ build tools are required.\n\n\u003C\u002Fdetails>\n\n---\n\n\u003Cdetails>\n\u003Csummary>\u003Cstrong> npm (cross-platform: Windows, macOS, Linux)\u003C\u002Fstrong>\u003C\u002Fsummary>\n\n```bash\n# Install globally\nsudo npm install -g open-mythos-2\n\n# Download the model (~4.5 GB, first time only)\nmythos model download\n\n# Start chatting\nmythos\n```\n\n> **Note:** The npm package wraps the same setup and Python backend under the hood. **Node.js 18+** and **Python 3.10+** are required. The first `mythos` run will automatically set up the virtual environment and dependencies if they aren't already present. On Windows, make sure Python is in your PATH.\n\n\u003C\u002Fdetails>\n\n\u003Cdetails>\n\u003Csummary>\u003Cstrong> pnpm (cross-platform: Windows, macOS, Linux)\u003C\u002Fstrong>\u003C\u002Fsummary>\n\n```bash\n# Install globally\npnpm install -g open-mythos-2\n\n# Download the model (~4.5 GB, first time only)\nmythos model download\n\n# Start chatting\nmythos\n```\n\n> **Note:** [pnpm](https:\u002F\u002Fpnpm.io\u002F) is a fast, disk-efficient package manager. The package wraps the same setup and Python backend under the hood. Node.js 18+ and Python 3.10+ are required. The first run will automatically set up the virtual environment and dependencies if they aren't already present.\n\n\u003C\u002Fdetails>\n\n\u003Cdetails>\n\u003Csummary>\u003Cstrong> pipx (Python users — cross-platform)\u003C\u002Fstrong>\u003C\u002Fsummary>\n\n```bash\n# Install with pipx (isolated environment, no venv management needed)\npipx install mythos-sentinel\n\n# Or install from source\ngit clone https:\u002F\u002Fgithub.com\u002Fcreatorofsomethingthatisgood\u002FOpen-Mythos-2.git\ncd Open-Mythos-2\npipx install .\n\n# Download the model (~4.5 GB, first time only)\nmythos model download\n\n# Start chatting\nmythos\n```\n\n> **Note:** [pipx](https:\u002F\u002Fpipx.pypa.io\u002F) installs Python CLI tools into isolated virtualenvs. You'll still need the C++ build tools (see platform-specific sections above) for `llama-cpp-python` to compile with GPU support. If building from source, `pipx install .` uses `pyproject.toml`.\n\n\u003C\u002Fdetails>\n\n\u003Cdetails>\n\u003Csummary>\u003Cstrong> Docker (cross-platform — no local deps needed)\u003C\u002Fstrong>\u003C\u002Fsummary>\n\n```bash\n# Pull and run (CPU-only)\ndocker run -it --rm \\\n  -v mythos-data:\u002Froot\u002F.config\u002Fmythos \\\n  ghcr.io\u002Fcreatorofsomethingthatisgood\u002Fopen-mythos-2:latest\n\n# With NVIDIA GPU acceleration\ndocker run -it --rm --gpus all \\\n  -v mythos-data:\u002Froot\u002F.config\u002Fmythos \\\n  ghcr.io\u002Fcreatorofsomethingthatisgood\u002Fopen-mythos-2:latest-cuda\n\n# Download model first (one-time)\ndocker run -it --rm \\\n  -v mythos-data:\u002Froot\u002F.config\u002Fmythos \\\n  ghcr.io\u002Fcreatorofsomethingthatisgood\u002Fopen-mythos-2:latest \\\n  mythos model download\n\n# Web UI\ndocker run -it --rm -p 7860:7860 \\\n  -v mythos-data:\u002Froot\u002F.config\u002Fmythos \\\n  ghcr.io\u002Fcreatorofsomethingthatisgood\u002Fopen-mythos-2:latest \\\n  mythos web --port 7860\n```\n\n> **Note:** Docker avoids the need for local C++ build tools. The CUDA image requires the [NVIDIA Container Toolkit](https:\u002F\u002Fdocs.nvidia.com\u002Fdatacenter\u002Fcloud-native\u002Fcontainer-toolkit\u002Finstall-guide.html). Model data persists in the `mythos-data` Docker volume.\n\n\u003C\u002Fdetails>\n\n\u003Cdetails>\n\u003Csummary>\u003Cstrong>⚡ Quick comparison\u003C\u002Fstrong>\u003C\u002Fsummary>\n\n| Method | Platforms | GPU | Best for |\n|--------|-----------|-----|----------|\n| **Git Clone** | Linux, macOS, Windows | Vulkan \u002F Metal \u002F CUDA | Full control, offline setups |\n| **npm \u002F pnpm** | Linux, macOS, Windows | Auto-detect | JS devs, quick install |\n| **pipx** | Linux, macOS, Windows | Manual CMAKE_ARGS | Python users |\n| **Docker** | Linux, macOS, Windows | CUDA (NVIDIA toolkit) | Reproducible, no local deps |\n| **Homebrew** | macOS only | Metal (auto) | Mac users who prefer brew |\n\n\u003C\u002Fdetails>\n\n\n---\n\n##  RML — Reinforcement Machine Learning\n\nRML is a feedback-driven self-improvement loop that learns your preferences over time and adapts Mythos to match. It does **two things automatically**:\n\n### 1. Edits the System Prompt\n\nWhen a category accumulates enough negative feedback, RML injects a concrete behavioral hint directly into the system prompt. These learned hints tell Mythos how to adjust its style — for example:\n\n- **Accuracy** hint: *\"Prioritize accuracy over creativity. Double-check facts. If unsure, say so.\"*\n- **Conciseness** hint: *\"Be concise. Get to the answer quickly; elaborate only when asked.\"*\n- **Clarity** hint: *\"Use headers, bullet points, short paragraphs. Avoid jargon without explanation.\"*\n- **Code quality** hint: *\"Write production-quality code: type hints, docstrings, error handling.\"*\n- **Security** hint: *\"Apply security best practices: input validation, no hardcoded secrets.\"*\n- **Completeness** hint: *\"Be thorough. Address all parts of the question. Don't skip edge cases.\"*\n\nHints are **removed automatically** when the category's score recovers — so the system prompt is always a live reflection of what you actually prefer.\n\n### 2. Adjusts Generation Parameters\n\nRML also tweaks `temperature`, `top_p`, and `repeat_penalty` behind the scenes:\n\n- High accept rate → temperature nudges **up** (more creative\u002Fconfident)\n- High rejections\u002Fedits → temperature nudges **down** (more conservative\u002Fprecise)\n- Many interrupts → `repeat_penalty` increases slightly\n\nAll adjustments are bounded by `max_param_offset` (default 0.3) so nothing swings wildly.\n\n### Feedback Signals RML Collects\n\n| Signal | Strength | Trigger |\n|--------|----------|---------|\n| Explicit good | +2.0 | `\u002Frml good` |\n| Explicit bad | −2.0 | `\u002Frml bad` |\n| Implicit positive | +1.0 | You say \"thanks\", \"perfect\", \"works\", etc. |\n| Implicit negative | −1.0 | You say \"wrong\", \"try again\", \"mistake\", etc. |\n| Edit penalty | −1.0 | You rewrite Mythos' output |\n| Interrupt penalty | −0.5 | You Ctrl+C during generation |\n\n### Commands\n\n```\n\u002Frml on          Enable RML\n\u002Frml off         Disable RML\n\u002Frml good        Mark last response as good (explicit +2)\n\u002Frml bad         Mark last response as bad (explicit -2)\n\u002Frml stats       Show what RML has learned (scores, hints, param adjustments)\n\u002Frml reset       Wipe all learned preferences and start fresh\n```\n\n### Config (`config.yaml`)\n\n```yaml\nrml:\n  enabled: false          # Turn on with \u002Frml on in chat\n  learning_rate: 0.05     # 0.01 = slow, 0.2 = fast\n  max_param_offset: 0.3   # Max drift from base temperature\u002Ftop_p\n  hint_threshold: 3.0     # Negative score before a hint is injected\n```\n\nPreferences persist in `~\u002F.config\u002Fmythos\u002Frml_preferences.json` across sessions.\n\n---\n\n## 📜 Command Reference\n\n### CLI Commands (terminal)\n\n| Command | Description |\n|---------|-------------|\n| `mythos` | Launch chat (default) |\n| `mythos chat` | Same as above (local, recommended) |\n| `mythos cloud` | Chat via cloud API (OpenAI-compatible) |\n| `mythos cloud set-key \u003Ckey> --provider nvidia` | Save API key with provider (nvidia, openai, together, groq) |\n| `mythos cloud status` | Show cloud configuration |\n| `mythos cloud clear` | Remove cloud API key |\n| `mythos chat --config \u003Cpath>` | Use a custom config file |\n| `mythos web` | Launch web UI (Gradio) |\n| `mythos web --port 8080 --share` | Custom port + public link |\n| `mythos init` | First-time setup |\n| `mythos status` | Show config & model status |\n| `mythos models` | List available GGUF models |\n| `mythos config show` | Display full resolved configuration |\n| `mythos doctor` | Diagnose setup issues & dependencies |\n| `mythos sessions` | List saved session summaries |\n| `mythos sessions -n 5` | Show last 5 sessions |\n| `mythos history` | List saved conversations |\n| `mythos history -n 10` | Show last 10 conversations |\n| `mythos scan` | Instant static security analysis |\n| `mythos scan --deep` | AI-powered audit (needs model) |\n| `mythos fix --path .` | Auto-fix safe patterns (dry-run) |\n| `mythos fix --path . --apply` | Apply fixes to disk |\n| `mythos path add ~\u002Fsrc` | Register a codebase |\n| `mythos path list` | List registered paths |\n| `mythos path remove \u003Ctarget>` | Remove a registered path |\n| `mythos model download` | Download default GGUF model |\n| `mythos update` | Pull latest from GitHub |\n\n### In-Chat Slash Commands\n\n| Command | Description |\n|---------|-------------|\n| `\u002Fhelp` | Show available commands |\n| `\u002Fconfig` | Show current configuration |\n| `\u002Fversion` | Show Mythos version and model info |\n| `\u002Ftokens` | Show token\u002Fgeneration stats |\n| `\u002Ftopp \u003C0.0-1.0>` | Set top-p (nucleus sampling) |\n| `\u002Ftopk \u003Cint>` | Set top-k sampling |\n| `\u002Freppen \u003Cfloat>` | Set repeat penalty |\n| `\u002Fmaxtokens \u003Cint>` | Set max tokens per generation |\n| `\u002Ftemp \u003C0.0-2.0>` | Set temperature |\n| `\u002Fpersona \u003Cname>` | Switch persona preset |\n| `\u002Fcompact` | Compact conversation context |\n| `\u002Fcopy` | Copy last response to clipboard |\n| `\u002Frename \u003Cname>` | Rename current conversation |\n| `\u002Fexport` | Export conversation as text |\n| `\u002Fdump` | Dump raw conversation JSON |\n| `\u002Fwc` | Word\u002Fchar\u002Ftoken count of conversation |\n| `\u002Fsave` | Save conversation |\n| `\u002Fsummary` | Generate and save a summary |\n| `\u002Frml on\\|off` | Enable\u002Fdisable RML |\n| `\u002Frml good\\|bad` | Mark last response |\n| `\u002Frml stats` | Show RML learning stats |\n| `\u002Frml reset` | Reset RML preferences |\n| `\u002Fquit` | Exit the chat |\n\n---\n\n## Voice Input (Whisper + AMD GPU)\n\nSpeak to Mythos instead of typing. Uses [whisper.cpp](https:\u002F\u002Fgithub.com\u002Fggerganov\u002Fwhisper.cpp) with the **Vulkan backend** for GPU-accelerated transcription on AMD GPUs (RDNA and newer).\n\n### Setup\n\n```bash\n# 1. Install whisper.cpp with Vulkan support (one-time, ~5 min)\n.\u002Fscripts\u002Finstall_whisper.sh          # base.en model (fast, 142 MB)\n.\u002Fscripts\u002Finstall_whisper.sh small    # small.en (better accuracy, 466 MB)\n```\n\n### Usage\n\n```bash\n# Enable voice mode\n\u002Fvoice on\n\n# Quick record (works even with voice off)\n\u002Frec\n\n# Press 'v' at an empty prompt to start recording, Enter to stop\nv\n(speak, then press Enter)\n\n# Check status\n\u002Fvoice\n\n# Disable\n\u002Fvoice off\n```\n\n### Config (`config.yaml`)\n\n```yaml\nvoice:\n  enabled: false\n  whisper_bin: \"whisper-cli\"          # or full path\n  model: \"models\u002Fggml-base.en.bin\"   # GGML model file\n  language: \"en\"                      # \"en\" or \"auto\"\n  push_to_talk: true\n  max_duration: 30                    # seconds\n```\n\n### AMD GPU Notes\n\n- **Vulkan backend** (recommended): works on all AMD GPUs with Vulkan 1.2+ (RX 6000\u002F7000\u002F9000 series)\n- **HIP\u002FROCm backend**: for supported cards only — build with `cmake -DGGML_HIP=1 -DAMDGPU_TARGETS=gfx1100`\n- Consumer GPUs (RX series) should use Vulkan — ROCm is mainly for MI-series datacenter GPUs\n\n---\n\n## Roadmap\n\n- [ ] Multi-model support (swap between GGUF models in-chat)\n- [ ] MCP (Model Context Protocol) server integration\n- [ ] LoRA fine-tuning from within the CLI\n- [ ] RAG over local documents\n- [ ] Plugin system for community extensions\n- [x] Voice input via Whisper (whisper.cpp, AMD Vulkan GPU support)\n- [ ] Voice output via Coqui TTS\n\n---\n\n## Contributing\n\nWe welcome contributions of all kinds — bug fixes, features, docs, and ideas.\n\n1. **Fork** the repo\n2. **Create a branch**: `git checkout -b feature\u002Fyour-idea`\n3. **Commit your changes**: `git commit -m \"Add your idea\"`\n4. **Push**: `git push origin feature\u002Fyour-idea`\n5. **Open a Pull Request**\n\nSee [Issues](https:\u002F\u002Fgithub.com\u002Fcreatorofsomethingthatisgood\u002FOpen-Mythos-2\u002Fissues) for open tasks. Feel free to open a new issue for bugs, feature requests, or questions.\n\n---\n\n\u003Cdiv align=\"center\">\n\n**If OpenMythos-2 speaks to you, leave a star — it helps others find the oracle.**\n\n[![Star History Chart](https:\u002F\u002Fapi.star-history.com\u002Fsvg?repos=creatorofsomethingthatisgood\u002FOpen-Mythos-2&type=Date)](https:\u002F\u002Fstar-history.com\u002F#creatorofsomethingthatisgood\u002FOpen-Mythos-2&Date)\n\n[![Stargazers repo roster for @creatorofsomethingthatisgood\u002FOpen-Mythos-2](https:\u002F\u002Freporoster.com\u002Fstars\u002Fcreatorofsomethingthatisgood\u002FOpen-Mythos-2)](https:\u002F\u002Fgithub.com\u002Fcreatorofsomethingthatisgood\u002FOpen-Mythos-2\u002Fstargazers)\n[![Forkers repo roster for @creatorofsomethingthatisgood\u002FOpen-Mythos-2](https:\u002F\u002Freporoster.com\u002Fforks\u002Fcreatorofsomethingthatisgood\u002FOpen-Mythos-2)](https:\u002F\u002Fgithub.com\u002Fcreatorofsomethingthatisgood\u002FOpen-Mythos-2\u002Fnetwork\u002Fmembers)\n\u003C\u002Fdiv>\n\n\nDisclaimer these answers may not be true by the model.\n","OpenMythos-2 是一个旨在通过聊天界面改进原始开放神话系统的项目。其核心功能是提供完全本地化的、无需API和云服务支持的人工智能体验，所有计算均在用户的终端上完成，确保了数据隐私。该项目使用Python语言编写，具有快速响应的特点，并且不依赖任何外部服务。它非常适合那些重视隐私安全、希望在没有网络连接的情况下也能利用AI技术的用户或开发者。无论是日常对话还是特定任务辅助，OpenMythos-2都能作为一个强大的工具被应用于个人电脑环境中。",2,"2026-06-11 04:01:48","CREATED_QUERY"]