{"id":48824521,"url":"https://github.com/terraphim/terraphim-ai-gpu-experimental","last_synced_at":"2026-04-14T17:30:55.814Z","repository":{"id":343523959,"uuid":"1178065900","full_name":"terraphim/terraphim-ai-gpu-experimental","owner":"terraphim","description":"Experimental GPU-accelerated desktop application using GPUI (Zed editor framework) - forked from terraphim-ai PR #461","archived":false,"fork":false,"pushed_at":"2026-03-10T17:30:23.000Z","size":335501,"stargazers_count":0,"open_issues_count":18,"forks_count":0,"subscribers_count":0,"default_branch":"main","last_synced_at":"2026-03-10T23:05:59.149Z","etag":null,"topics":[],"latest_commit_sha":null,"homepage":null,"language":"Rust","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"apache-2.0","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/terraphim.png","metadata":{"files":{"readme":"README.md","changelog":"historys_plan.md","contributing":"CONTRIBUTING.md","funding":null,"license":"LICENSE-Apache-2.0","code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null,"zenodo":null,"notice":null,"maintainers":null,"copyright":null,"agents":"AGENTS.md","dco":null,"cla":null}},"created_at":"2026-03-10T16:47:33.000Z","updated_at":"2026-03-10T16:57:21.000Z","dependencies_parsed_at":null,"dependency_job_id":null,"html_url":"https://github.com/terraphim/terraphim-ai-gpu-experimental","commit_stats":null,"previous_names":["terraphim/terraphim-ai-gpu-experimental"],"tags_count":56,"template":false,"template_full_name":null,"purl":"pkg:github/terraphim/terraphim-ai-gpu-experimental","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/terraphim%2Fterraphim-ai-gpu-experimental","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/terraphim%2Fterraphim-ai-gpu-experimental/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/terraphim%2Fterraphim-ai-gpu-experimental/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/terraphim%2Fterraphim-ai-gpu-experimental/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/terraphim","download_url":"https://codeload.github.com/terraphim/terraphim-ai-gpu-experimental/tar.gz/refs/heads/main","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/terraphim%2Fterraphim-ai-gpu-experimental/sbom","scorecard":null,"host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":286080680,"owners_count":31808505,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2026-04-14T11:13:53.975Z","status":"ssl_error","status_checked_at":"2026-04-14T11:13:53.299Z","response_time":153,"last_error":"SSL_connect returned=1 errno=0 peeraddr=140.82.121.5:443 state=error: unexpected eof while reading","robots_txt_status":"success","robots_txt_updated_at":"2025-07-24T06:49:26.215Z","robots_txt_url":"https://github.com/robots.txt","online":false,"can_crawl_api":true,"host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":[],"created_at":"2026-04-14T17:30:54.470Z","updated_at":"2026-04-14T17:30:55.793Z","avatar_url":"https://github.com/terraphim.png","language":"Rust","funding_links":[],"categories":[],"sub_categories":[],"readme":"# Terraphim AI Assistant\n\n[![Discord](https://img.shields.io/discord/852545081613615144?label=Discord\u0026logo=Discord)](https://discord.gg/VPJXB6BGuY)\n[![Discourse](https://img.shields.io/discourse/users?server=https%3A%2F%2Fterraphim.discourse.group)](https://terraphim.discourse.group)\n\nTerraphim is a privacy-first AI assistant that works for you under your complete control and is fully deterministic.\n\nYou can use it as a local search engine, configured to search for different types of content on StackOverflow, GitHub, and the local filesystem using a predefined folder, which includes your Markdown files.\n\nTerraphim operates on local infrastructure and works exclusively for the owner's benefit.\n\nhttps://github.com/terraphim/terraphim-ai/assets/175809/59c74652-bab4-45b2-99aa-1c0c9b90196b\n\n\n## Why Terraphim?\n\nThere are growing concerns about the privacy of data and the sharing of individuals' data across an ever-growing list of services, some of which have questionable data ethics policies. \u003csup\u003e[1],[2],[3],[4]\u003c/sup\u003e\n\n**Individuals struggle to find relevant information in different knowledge repositories:**\n\n- Personal ones like Roam Research, Obsidian, Coda, and Notion.\n- Team-focused ones like Jira, Confluence, and SharePoint.\n- Public sources such as StackOverflow and GitHub.\n\nTerraphim aims to bridge this gap by providing a privacy-first AI assistant that operates locally on the user's hardware, enabling seamless access to various knowledge repositories without compromising privacy. With Terraphim, users can efficiently search personal, team-focused, and public knowledge sources, ensuring that their data remains under their control at all times.\n\n[1]: https://www.coveo.com/en/resources/reports/relevance-report-workplace\n[2]: https://cottrillresearch.com/various-survey-statistics-workers-spend-too-much-time-searching-for-information/\n[3]: https://www.forbes.com/sites/forbestechcouncil/2019/12/17/reality-check-still-spending-more-time-gathering-instead-of-analyzing/\n[4]: https://www.theatlantic.com/technology/archive/2021/06/the-internet-is-a-collective-hallucination/619320/\n\n## Getting Started\n\n### 🚀 Quick Install (Recommended)\n\n#### Option 1: Docker (Easiest)\n```bash\n# Automated Docker installation\ncurl -fsSL https://raw.githubusercontent.com/terraphim/terraphim-ai/main/release/v0.2.3/docker-run.sh | bash\n```\n\n#### Option 2: Binary Installation\n```bash\n# Automated source installation\ncurl -fsSL https://raw.githubusercontent.com/terraphim/terraphim-ai/main/release/v0.2.3/install.sh | bash\n```\n\n### 📚 Detailed Installation\nFor detailed installation instructions, see our [Installation Guide](https://github.com/terraphim/terraphim-ai/blob/main/release/v0.2.3/README.md).\n\n### 🛠️ Development Setup\n\n1. **Clone the repository**:\n   ```bash\n   git clone https://github.com/terraphim/terraphim-ai.git\n   cd terraphim-ai\n   ```\n\n2. **Set up development environment**:\n   ```bash\n   # Install Git hooks for code quality (optional but recommended)\n   ./scripts/install-hooks.sh\n   ```\n\n3. **Start the backend server**:\n   ```bash\n   cargo run\n   ```\n   This starts an API endpoint for indexing and querying documents.\n\n4. **Run the frontend** (choose one):\n\n   **Web Frontend:**\n   ```bash\n   cd desktop\n   yarn install\n   yarn run dev\n   ```\n\n   **Desktop App (Tauri):**\n   ```bash\n   cd desktop\n   yarn install\n   yarn run tauri dev\n   ```\n\n   **Terminal Interface (TUI):**\n   ```bash\n   # Build with all features (recommended)\n   cargo build -p terraphim_tui --features repl-full --release\n   ./target/release/terraphim-tui\n\n   # Or run minimal version\n   cargo run --bin terraphim-tui\n   ```\n\n(See the [desktop README](desktop/README.md), [TUI documentation](docs/tui-usage.md), and [development setup guide](docs/src/development-setup.md) for more details.)\n\n## Terminal User Interface (TUI)\n\nTerraphim includes a comprehensive TUI that provides both interactive REPL functionality and CLI commands for advanced operations:\n\n### Key Features\n\n- **🤖 AI Chat Integration**: OpenRouter and Ollama support for intelligent conversations\n- **⚡ Sub-2 Second VM Boot**: Advanced VM optimization system with sub-500ms allocation\n- **🖥️ Enhanced VM Management**: Firecracker microVM pools with intelligent allocation\n- **📝 Markdown Command System**: Extensible commands defined in YAML frontmatter\n- **🔒 Security-First Execution**: Three execution modes (Local, Firecracker, Hybrid) with knowledge graph validation\n- **🌐 Web Operations**: Secure web requests through VM sandboxing\n- **📁 File Operations**: Semantic file analysis and intelligent content management\n- **🔍 Knowledge Graph**: Interactive rolegraph visualization and navigation\n- **⚙️ Configuration**: Real-time role and configuration management\n\n### Quick Start\n\n```bash\n# Build with all features\ncargo build -p terraphim_tui --features repl-full --release\n\n# Launch interactive REPL\n./target/release/terraphim-tui\n\n# Available REPL commands:\n/help           # Show all commands\n/search \"query\" # Semantic search\n/chat \"message\" # AI conversation\n/commands list  # List available markdown commands\n/deploy staging # Execute deployment (Firecracker mode)\n/search \"TODO\"  # Execute search command (Local mode)\n/vm list        # VM management with sub-2s boot\n/web get URL    # Web operations\n/file search    # Semantic file operations\n```\n\nFor detailed documentation, see [TUI Usage Guide](docs/tui-usage.md).\n\n## Terminology\n\nWhen configuring or working on Terraphim, you will encounter the following\nterms and concepts:\n\n- **Haystack**: A data source that Terraphim can search through. For example, this\n  could be a folder on your computer, a Notion workspace, or your email account.\n- **Knowledge Graph**: A structured graph of information created from a\n  haystack, where nodes represent entities and edges represent relationships\n  between them.\n- **Profile**: An endpoint for persisting user data (e.g. Amazon S3, sled, or\n  rocksdb).\n- **Role**: A role is a set of settings that define the default behavior of the\n  AI assistant. For example, a developer role will search for code-related\n  content, while a \"father\" role might search for parenting-related content. Each\n  Terraphim role has its own separate knowledge graph that contains relevant\n  concepts, with all synonyms.\n- **Rolegraph**: A structure for ingesting documents into Terraphim. It is a knowledge\n  graph that uses a scoring function (an Aho-Corasick automata build from the\n  knowledge graph) for ranking results.\n\n## Why \"Terraphim\"?\n\nThe term is originally taken from the [Relict series][relict] of science fiction\nnovels by [Vasiliy Golovachev](https://en.wikipedia.org/wiki/Vasili_Golovachov).\nTerraphim is an artificial intelligence living inside a spacesuit (part of an\nexocortex), or inside your house or vehicle, and it is designed to help you with\nyour tasks. You can carry it around with you.\nSimilar entities now common in science fiction, for example Destiny 2 has entity called [Ghost][ghost].\n\n\nOr in Star Wars Jedi Survivor there is an AI assistant [BD-1][bd-1].\n\nThe compactness and mobility of such AI assistant drives the [[Design Decisions]] of Terraphim.\n\n[bd-1]: https://starwars.fandom.com/wiki/BD-1\n[ghost]: https://www.destinypedia.com/Ghost\n[relict]: https://www.goodreads.com/en/book/show/196710046\n\nTerraphim is a trademark registered in the UK, US and internationally (WIPO). All other trademarks mentioned above are the property of their respective owners.\n\n## Configuration\n\n### Storage Backends\n\nTerraphim supports multiple storage backends for different deployment scenarios:\n\n#### Local Development (Default)\nThe system uses local storage backends by default, requiring no additional configuration:\n- **Memory**: In-memory storage for testing\n- **DashMap**: High-performance concurrent storage\n- **SQLite**: Local database storage\n- **ReDB**: Embedded key-value database\n\n#### Cloud Storage (Optional)\nFor production deployments, you can optionally enable cloud storage:\n\n##### AWS S3 Configuration\nTo use AWS S3 storage, set the following environment variables:\n```bash\nexport AWS_ACCESS_KEY_ID=\"your-access-key\"\nexport AWS_SECRET_ACCESS_KEY=\"your-secret-key\"\nexport TERRAPHIM_PROFILE_S3_REGION=\"us-east-1\"\nexport TERRAPHIM_PROFILE_S3_ENDPOINT=\"https://s3.amazonaws.com/\"\n```\n\n**Note**: AWS credentials are completely optional. The system will automatically fall back to local storage if AWS credentials are not available, ensuring local development works without any cloud dependencies.\n\n### Environment Variables\n- `TERRAPHIM_SETTINGS_PATH`: Override the settings file path\n- `TERRAPHIM_DATA_PATH`: Set the data directory location\n- `LOG_LEVEL`: Set logging verbosity (debug, info, warn, error)\n\n## Installation Methods\n\n### For End Users\n\n#### Homebrew (macOS/Linux)\n```bash\nbrew install terraphim/terraphim-ai/terraphim-ai\n```\nThis installs the server, TUI, and desktop app (macOS only).\n\n#### Debian/Ubuntu\n```bash\n# Download from GitHub releases\nsudo dpkg -i terraphim-server_*.deb\nsudo dpkg -i terraphim-tui_*.deb\nsudo dpkg -i terraphim-ai-desktop_*.deb\n```\n\n#### Docker\n```bash\ndocker run ghcr.io/terraphim/terraphim-server:latest\n```\n\n#### Direct Download\nDownload pre-built binaries from [GitHub Releases](https://github.com/terraphim/terraphim-ai/releases).\n\n### Development Setup\n\nFor development, see our comprehensive [Development Setup Guide](docs/src/development-setup.md) which covers:\n- Code quality tools and pre-commit hooks\n- Multiple installation options (no Python required!)\n- IDE integration and troubleshooting\n\n## Claude Code Integration\n\nTerraphim provides seamless integration with Claude Code through multiple approaches, enabling intelligent text replacement and codebase quality evaluation.\n\n### 🔄 Text Replacement (Hooks \u0026 Skills)\n\nUse Terraphim's knowledge graph capabilities to automatically replace text patterns in your development workflow:\n\n**Claude Code Hooks** - Automatic, transparent replacements:\n```bash\n# Example: Automatically replace npm with bun\necho \"npm install\" | terraphim-tui replace\n# Output: bun_install\n```\n\n**Claude Skills** - Context-aware, conversational assistance:\n- Works across all Claude platforms\n- Provides explanations and reasoning\n- Progressive disclosure of functionality\n\n**Examples:**\n- Package manager enforcement (npm/yarn/pnpm → bun)\n- Attribution replacement (Claude Code → Terraphim AI)\n- Custom domain-specific replacements\n\n📖 **Complete Guide**: [examples/TERRAPHIM_CLAUDE_INTEGRATION.md](examples/TERRAPHIM_CLAUDE_INTEGRATION.md)\n\n### 📊 Codebase Quality Evaluation\n\nEvaluate whether AI agents (Claude Code, GitHub Copilot, autonomous agents) improve or deteriorate your codebase using deterministic, knowledge graph-based assessment:\n\n**Key Features:**\n- **Deterministic**: Aho-Corasick automata for consistent scoring\n- **Privacy-First**: All evaluation runs locally\n- **Multi-Dimensional**: Security, performance, quality perspectives\n- **CI/CD Ready**: Automated quality gates with exit codes\n\n**Quick Start:**\n```bash\ncd examples/codebase-evaluation\n./scripts/evaluate-ai-agent.sh /path/to/codebase\n\n# Generates verdict:\n# ✅ IMPROVEMENT: The AI agent improved the codebase quality.\n# - Improved metrics: 3\n# - Deteriorated metrics: 0\n```\n\n**Evaluation Metrics:**\n- Clippy warnings, anti-patterns, TODO counts\n- Knowledge graph density and semantic matches\n- Test pass rates and code coverage\n- Custom domain-specific patterns\n\n**Use Cases:**\n1. Evaluate PRs from AI agents before merge\n2. Continuous quality monitoring in CI/CD pipelines\n3. Historical trend analysis across evaluations\n4. Multi-role evaluation (security + performance + quality)\n\n📖 **Complete Documentation**: [examples/codebase-evaluation/](examples/codebase-evaluation/)\n\n**Example GitHub Actions Integration:**\n```yaml\n- name: Baseline evaluation\n  run: ./scripts/baseline-evaluation.sh ${{ github.workspace }}\n- name: Apply AI changes\n  run: # Your AI agent step\n- name: Post-change evaluation\n  run: ./scripts/post-evaluation.sh ${{ github.workspace }}\n- name: Generate verdict (fails on deterioration)\n  run: ./scripts/compare-evaluations.sh\n```\n\n## Contributing\n\nWe welcome contributions! Here's how to get started:\n\n1. **Read our guides**:\n   - [Contributing guide](CONTRIBUTING.md)\n   - [Development setup](docs/src/development-setup.md)\n\n2. **Set up your environment**:\n   ```bash\n   git clone https://github.com/terraphim/terraphim-ai.git\n   cd terraphim-ai\n   ./scripts/install-hooks.sh  # Sets up code quality tools\n   ```\n\n3. **Code quality standards**:\n   - All commits must follow [Conventional Commits](https://www.conventionalcommits.org/)\n   - Rust code is automatically formatted with `cargo fmt`\n   - JavaScript/TypeScript uses [Biome](https://biomejs.dev/) for linting and formatting\n   - No secrets or large files allowed in commits\n\n4. **Make your changes**:\n   - Create a feature branch: `git checkout -b feat/your-feature`\n   - Make your changes with proper tests\n   - Commit with conventional format: `git commit -m \"feat: add amazing feature\"`\n   - Push and create a Pull Request\n\n### Dependency Management\n\n**Important**: Some dependencies are pinned to specific versions to ensure compatibility:\n\n- `wiremock = \"0.6.4\"` - Version 0.6.5 uses unstable Rust features requiring nightly compiler\n- `schemars = \"0.8.22\"` - Version 1.0+ introduces breaking API changes\n- `thiserror = \"1.0.x\"` - Version 2.0+ requires code migration for breaking changes\n\nThese constraints are enforced in `.github/dependabot.yml` to prevent automatic updates that would break CI. Future upgrades require manual code migration work.\n\n### Contributors are awesome\n\u003ca href=\"https://github.com/terraphim/terraphim-ai/graphs/contributors\"\u003e\n  \u003cimg src=\"https://contrib.rocks/image?repo=terraphim/terraphim-ai\" /\u003e\n\u003c/a\u003e\n\n\n\n## License\n\nThis project is licensed under the [Apache license](LICENSE).\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fterraphim%2Fterraphim-ai-gpu-experimental","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fterraphim%2Fterraphim-ai-gpu-experimental","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fterraphim%2Fterraphim-ai-gpu-experimental/lists"}