[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"project-74691":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":14,"contributorsCount":15,"subscribersCount":15,"size":15,"stars1d":15,"stars7d":16,"stars30d":17,"stars90d":15,"forks30d":15,"starsTrendScore":15,"compositeScore":18,"rankGlobal":10,"rankLanguage":10,"license":19,"archived":20,"fork":20,"defaultBranch":21,"hasWiki":20,"hasPages":22,"topics":23,"createdAt":10,"pushedAt":10,"updatedAt":29,"readmeContent":30,"aiSummary":31,"trendingCount":15,"starSnapshotCount":15,"syncStatus":32,"lastSyncTime":33,"discoverSource":34},74691,"localgpt","localgpt-app\u002Flocalgpt","localgpt-app","Local AI assistant, dreaming explorable worlds.","https:\u002F\u002Flocalgpt.app\u002F",null,"Rust",1103,101,6,0,3,4,55.93,"Apache License 2.0",false,"main",true,[24,25,26,27,28],"agent","ai","bevy","rust","world","2026-06-12 04:01:15","\n# \u003Cimg src=\"https:\u002F\u002Flocalgpt.app\u002Flogo\u002Flocalgpt-icon-app.svg\" width=\"50\" height=\"50\" alt=\"LocalGPT\" \u002F> LocalGPT\n\n[![License](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002Flicense-Apache%202.0-blue.svg)](https:\u002F\u002Fgithub.com\u002Flocalgpt-app\u002Flocalgpt#license)\n[![Crates.io](https:\u002F\u002Fimg.shields.io\u002Fcrates\u002Fv\u002Flocalgpt.svg)](https:\u002F\u002Fcrates.io\u002Fcrates\u002Flocalgpt)\n[![Downloads](https:\u002F\u002Fimg.shields.io\u002Fcrates\u002Fd\u002Flocalgpt.svg)](https:\u002F\u002Fcrates.io\u002Fcrates\u002Flocalgpt)\n[![Docs](https:\u002F\u002Fdocs.rs\u002Flocalgpt\u002Fbadge.svg)](https:\u002F\u002Fdocs.rs\u002Flocalgpt\u002Flatest\u002Flocalgpt)\n[![CI](https:\u002F\u002Fgithub.com\u002Flocalgpt-app\u002Flocalgpt\u002Fworkflows\u002FCI\u002Fbadge.svg)](https:\u002F\u002Fgithub.com\u002Flocalgpt-app\u002Flocalgpt\u002Factions)\n[![Discord](https:\u002F\u002Fimg.shields.io\u002Fdiscord\u002F691052431525675048.svg?label=&logo=discord&logoColor=ffffff&color=7389D8&labelColor=6A7EC2)](https:\u002F\u002Fdiscord.gg\u002FspKRr6mRyp)\n\nBuild explorable 3D worlds with natural language — geometry, materials, lighting, audio, and behaviors. Open source, runs locally.\n\n[![LocalGPT Gen Demo](https:\u002F\u002Fimg.youtube.com\u002Fvi\u002FR__tg7YY0T8\u002Fmaxresdefault.jpg)](https:\u002F\u002Fwww.youtube.com\u002Fwatch?v=R__tg7YY0T8)\n\n## Install\n\n### From crates.io (for users)\n\nIf you just want to run LocalGPT — no source checkout needed:\n\n```bash\n# World Building\ncargo install localgpt-gen\n\n# AI Assistant (chat, memory, daemon)\ncargo install localgpt\n```\n\n### From source (for developers)\n\nIf you've cloned the repository and want to hack on the code, use `cargo run` to iterate without installing:\n\n```bash\ngit clone https:\u002F\u002Fgithub.com\u002Flocalgpt-app\u002Flocalgpt.git\ncd localgpt\n\n# World Building\ncargo run -p localgpt-gen -- \"Create a desert scene with pyramids\"\n\n# AI Assistant\ncargo run -- chat\ncargo run -- daemon start\n```\n\nFull options (feature flags, headless builds, Docker): see [Installation](https:\u002F\u002Flocalgpt.app\u002Fdocs\u002Finstallation).\n\n## \u003Cimg src=\"https:\u002F\u002Flocalgpt.app\u002Flogo\u002Flocalgpt-icon.svg\" width=\"32\" height=\"32\" alt=\"LocalGPT\" \u002F> Gen Mode (World Building)\n\n`localgpt-gen` is a standalone binary for AI-driven 3D world creation with the Bevy game engine.\n\n```bash\n# Start interactive mode\nlocalgpt-gen\n\n# Start with an initial prompt\nlocalgpt-gen \"Create a desert scene with pyramids and a UFO hovering above\"\n\n# Load an existing scene\nlocalgpt-gen --scene .\u002Fworld.glb\n\n# Verbose logging\nlocalgpt-gen --verbose\n```\n\n### Features\n\n- **Parametric shapes** — box, sphere, cylinder, capsule, plane, torus, pyramid, tetrahedron, icosahedron, wedge\n- **PBR materials** — color, metalness, roughness, emissive, alpha, double-sided\n- **Lighting** — point, spot, directional lights with color and intensity\n- **Behaviors** — orbit, spin, bob, look_at, pulse, path_follow, bounce\n- **Audio** — ambient sounds (wind, rain, forest, ocean, cave) and spatial emitters\n- **Export** — glTF\u002FGLB, HTML (browser-viewable), screenshots\n- **World skills** — save\u002Fload complete worlds as reusable skills\n\n### Headless Mode & Experiment Queue\n\nQueue world experiments and generate without a window — overnight batch runs, CI pipelines, or scripted variations:\n\n```bash\n# Generate a single world (no window)\nlocalgpt-gen headless --prompt \"Build a cozy cabin in a snowy forest\"\n\n# With style hint\nlocalgpt-gen headless --prompt \"Village marketplace\" --style \"Studio Ghibli\"\n```\n\nThe memory system learns your creative style across sessions — palettes, lighting preferences, entity templates — and applies them automatically in future generations.\n\nFull docs: [Headless Mode & Experiment Queue](https:\u002F\u002Flocalgpt.app\u002Fdocs\u002Fgen\u002Fheadless)\n\n### MCP Server\n\nUse Gen from Claude Desktop, Codex Desktop, or any MCP-compatible editor (VS Code, Zed, Cursor):\n\n```bash\nlocalgpt-gen mcp-server\n```\n\nAdd to Claude Desktop config (`~\u002FLibrary\u002FApplication Support\u002FClaude\u002Fclaude_desktop_config.json` on macOS):\n\n```json\n{\n  \"mcpServers\": {\n    \"localgpt-gen\": {\n      \"command\": \"localgpt-gen\",\n      \"args\": [\"mcp-server\"]\n    }\n  }\n}\n```\n\nFor CLI tools (Claude CLI, Gemini CLI, Codex CLI), use `--connect` to route tool calls to your existing window. See [CLI Mode (MCP Relay)](https:\u002F\u002Flocalgpt.app\u002Fdocs\u002Fgen\u002Fcli-mode).\n\nFull docs: [LocalGPT Gen](https:\u002F\u002Flocalgpt.app\u002Fdocs\u002Fgen) | [MCP Server](https:\u002F\u002Flocalgpt.app\u002Fdocs\u002Fgen\u002Fmcp-server)\n\nBuilt something cool? Share on [Discord](https:\u002F\u002Fdiscord.gg\u002FspKRr6mRyp) or [YouTube](https:\u002F\u002Fwww.youtube.com\u002F@localgpt-gen)!\n\n---\n\n## AI Assistant\n\n`localgpt` is a local-first AI assistant with persistent memory, autonomous tasks, and multiple interfaces.\n\n```bash\n# Interactive chat\nlocalgpt chat\n\n# Single question\nlocalgpt ask \"What is the meaning of life?\"\n\n# Run as daemon with HTTP API and web UI\nlocalgpt daemon start\n```\n\n### Why LocalGPT?\n\n- **Single binary** — no Node.js, Docker, or Python required\n- **Local device focused** — runs entirely on your machine, your data stays yours\n- **Persistent memory** — markdown-based knowledge store with full-text and semantic search\n- **Hybrid web search** — native provider search passthrough plus client-side fallback\n- **Autonomous heartbeat** — delegate tasks and let it work in the background\n- **Multiple interfaces** — CLI, web UI, desktop GUI, Telegram bot\n- **Defense-in-depth security** — signed policy files, kernel-enforced sandbox, prompt injection defenses\n- **Multiple LLM providers** — LM Studio, Ollama, Anthropic, OpenAI, xAI, GLM, Vertex AI, CLI providers\n\n### How It Works\n\nLocalGPT uses XDG-compliant directories for config\u002Fdata\u002Fstate\u002Fcache. Run `localgpt paths` to see resolved paths.\n\nWorkspace memory layout:\n\n```\n\u003Cworkspace>\u002F\n├── MEMORY.md     # Long-term knowledge (auto-loaded each session)\n├── HEARTBEAT.md  # Autonomous task queue\n├── SOUL.md       # Personality and behavioral guidance\n└── knowledge\u002F    # Structured knowledge bank\n```\n\nFiles are indexed with SQLite FTS5 for keyword search and sqlite-vec for semantic search with local embeddings.\n\n### Configuration\n\nStored at `\u003Cconfig_dir>\u002Fconfig.toml`:\n\n**Local models via LM Studio** (no API key, fully offline):\n\n```toml\n[agent]\ndefault_model = \"openai\u002Fqwen\u002Fqwen3.5-35b-a3b\"\n\n[providers.openai]\napi_key = \"lm-studio\"\nbase_url = \"http:\u002F\u002F127.0.0.1:1234\u002Fv1\"\n```\n\n**Cloud providers** (Anthropic, OpenAI, etc.):\n\n```toml\n[agent]\ndefault_model = \"claude-cli\u002Fopus\"\n\n[providers.anthropic]\napi_key = \"${ANTHROPIC_API_KEY}\"\n```\n\nFull config reference: [`website\u002Fdocs\u002Fconfiguration.md`](website\u002Fdocs\u002Fconfiguration.md)\n\n### Security\n\n- **Kernel-enforced sandbox** — Landlock\u002Fseccomp on Linux, Seatbelt on macOS\n- **Signed policy files** — HMAC-SHA256 signed `LocalGPT.md` with tamper detection\n- **Prompt injection defenses** — marker stripping, pattern detection, content boundaries\n- **Audit chain** — hash-chained security event log\n\nSecurity docs: [`website\u002Fdocs\u002Fsandbox.md`](website\u002Fdocs\u002Fsandbox.md) | [`website\u002Fdocs\u002Flocalgpt.md`](website\u002Fdocs\u002Flocalgpt.md)\n\n### HTTP API\n\n| Endpoint | Description |\n|----------|-------------|\n| `GET \u002F` | Embedded web UI |\n| `POST \u002Fapi\u002Fchat` | Chat with assistant |\n| `POST \u002Fapi\u002Fchat\u002Fstream` | SSE streaming chat |\n| `GET \u002Fapi\u002Fmemory\u002Fsearch?q=\u003Cquery>` | Search memory |\n\nFull API reference: [`website\u002Fdocs\u002Fhttp-api.md`](website\u002Fdocs\u002Fhttp-api.md)\n\n### CLI Commands\n\n```bash\nlocalgpt chat                   # Interactive chat\nlocalgpt ask \"question\"         # Single question\nlocalgpt daemon start           # Start daemon\nlocalgpt memory search \"query\"  # Search memory\nlocalgpt config show            # Show config\nlocalgpt paths                  # Show resolved paths\n```\n\nFull CLI reference: [`website\u002Fdocs\u002Fcli-commands.md`](website\u002Fdocs\u002Fcli-commands.md)\n\n## Blog\n\n- [Explorable World as Agent Skill](https:\u002F\u002Flocalgpt.app\u002Fblog\u002Fworld-as-skill)\n- [Why I Built LocalGPT in 4 Nights](https:\u002F\u002Flocalgpt.app\u002Fblog\u002Fwhy-i-built-localgpt-in-4-nights)\n\n## Built With\n\nRust, Tokio, Axum, Bevy, SQLite (FTS5 + sqlite-vec), fastembed, eframe\n\n## Contributors\n\n\u003Ca href=\"https:\u002F\u002Fgithub.com\u002Flocalgpt-app\u002Flocalgpt\u002Fgraphs\u002Fcontributors\">\n  \u003Cimg src=\"https:\u002F\u002Fcontrib.rocks\u002Fimage?repo=localgpt-app\u002Flocalgpt\" \u002F>\n\u003C\u002Fa>\n\n## License\n\n[Apache-2.0](LICENSE)\n","LocalGPT 是一个基于自然语言生成可探索3D世界的本地AI助手。它利用Rust编写，并采用Bevy游戏引擎，支持创建包括几何体、材质、光照、音频和行为等在内的丰富场景元素。用户可以通过简单的命令行指令或交互模式来构建如沙漠金字塔等复杂环境。此外，LocalGPT还提供了无头模式与实验队列功能，便于批量处理及集成到CI\u002FCD流程中。该项目适用于需要快速原型设计或想要通过编程方式创造虚拟世界的开发者和设计师。",2,"2026-06-11 03:50:25","high_star"]