{"id":49418385,"url":"https://github.com/abd0r/artificial-neural-mesh-v0","last_synced_at":"2026-04-29T04:12:43.386Z","repository":{"id":331286514,"uuid":"1125945125","full_name":"Abd0r/Artificial-Neural-Mesh-V0","owner":"Abd0r","description":"A Multi Agent Reasoning System.","archived":false,"fork":false,"pushed_at":"2026-01-02T17:52:58.000Z","size":3372,"stargazers_count":0,"open_issues_count":0,"forks_count":0,"subscribers_count":0,"default_branch":"main","last_synced_at":"2026-03-01T22:00:24.919Z","etag":null,"topics":["ai","ai-agents","artificial-intelligence","artificial-neural-mesh","refiner","verifier","web-of-toughts"],"latest_commit_sha":null,"homepage":"","language":"Python","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"mit","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/Abd0r.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":"LICENSE","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":null,"dco":null,"cla":null}},"created_at":"2025-12-31T17:43:29.000Z","updated_at":"2026-02-27T14:25:03.000Z","dependencies_parsed_at":"2026-01-06T03:09:19.601Z","dependency_job_id":null,"html_url":"https://github.com/Abd0r/Artificial-Neural-Mesh-V0","commit_stats":null,"previous_names":["ra2157218-boop/artificial-neural-mesh-v0","abd0r/artificial-neural-mesh-v0"],"tags_count":1,"template":false,"template_full_name":null,"purl":"pkg:github/Abd0r/Artificial-Neural-Mesh-V0","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/Abd0r%2FArtificial-Neural-Mesh-V0","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/Abd0r%2FArtificial-Neural-Mesh-V0/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/Abd0r%2FArtificial-Neural-Mesh-V0/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/Abd0r%2FArtificial-Neural-Mesh-V0/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/Abd0r","download_url":"https://codeload.github.com/Abd0r/Artificial-Neural-Mesh-V0/tar.gz/refs/heads/main","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/Abd0r%2FArtificial-Neural-Mesh-V0/sbom","scorecard":null,"host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":286080680,"owners_count":32410085,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2026-04-29T03:46:11.172Z","status":"ssl_error","status_checked_at":"2026-04-29T03:37:55.317Z","response_time":110,"last_error":"SSL_connect returned=1 errno=0 peeraddr=140.82.121.6:443 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":["ai","ai-agents","artificial-intelligence","artificial-neural-mesh","refiner","verifier","web-of-toughts"],"created_at":"2026-04-29T04:12:40.774Z","updated_at":"2026-04-29T04:12:43.377Z","avatar_url":"https://github.com/Abd0r.png","language":"Python","funding_links":[],"categories":[],"sub_categories":[],"readme":"\u003cp align=\"center\"\u003e\n  \u003cimg src=\"docs/ANM-White-Logo.png\" alt=\"ANM Logo\" width=\"400\"\u003e\n\u003c/p\u003e\n\n\u003ch1 align=\"center\"\u003eArtificial Neural Mesh (ANM) V0-OpenSource\u003c/h1\u003e\n\n\u003cp align=\"center\"\u003e\n  \u003cimg src=\"https://img.shields.io/badge/Python-3.9--3.13-blue?style=flat\u0026logo=python\" alt=\"Python\" height=\"18\" style=\"border-radius: 5%;\"\u003e\n  \u003cimg src=\"https://img.shields.io/badge/License-MIT-green?style=flat\" alt=\"License\" height=\"18\" style=\"border-radius: 5%;\"\u003e\n  \u003cimg src=\"https://img.shields.io/badge/Status-Research%20Artifact-orange?style=flat\" alt=\"Status\" height=\"18\" style=\"border-radius: 5%;\"\u003e\n  \u003ca href=\"https://github.com/ra2157218-boop/Artificial-Neural-Mesh-V0/releases/tag/v0.1.0\"\u003e\u003cimg src=\"https://img.shields.io/badge/Release-v0.1.0-purple?style=flat\u0026logo=github\" alt=\"Release\" height=\"18\" style=\"border-radius: 5%;\"\u003e\u003c/a\u003e\n  \u003ca href=\"https://zenodo.org/records/18112435\"\u003e\u003cimg src=\"https://img.shields.io/badge/DOI-10.5281%2Fzenodo.18112435-blue?style=flat\u0026logo=zenodo\" alt=\"DOI\" height=\"18\" style=\"border-radius: 5%;\"\u003e\u003c/a\u003e\n  \u003ca href=\"https://huggingface.co/datasets/Abd0r/anm-v0-benchmark\"\u003e\u003cimg src=\"https://img.shields.io/badge/Dataset-HuggingFace-FFE066?style=flat\u0026logo=huggingface\" alt=\"HuggingFace Dataset\" height=\"18\" style=\"border-radius: 5%;\"\u003e\u003c/a\u003e\n  \u003ca href=\"https://x.com/SyedAbdurR2hman\"\u003e\u003cimg src=\"https://img.shields.io/badge/Author-@SyedAbdurR2hman-black?style=flat\u0026logo=x\" alt=\"Author/Updates\" height=\"18\" style=\"border-radius: 5%;\"\u003e\u003c/a\u003e\n\u003c/p\u003e\n\n\u003cp align=\"center\"\u003e\n  \u003cb\u003eA Multi-Agent AI System with Web-of-Thought Reasoning\u003c/b\u003e\n\u003c/p\u003e\n\n---\n\n## What is ANM?\n\n**Artificial Neural Mesh (ANM)** is an advanced multi-agent AI system that combines 12 specialized domain experts with a novel **Web-of-Thought (WoT)** reasoning engine. Unlike traditional single-model approaches, ANM routes queries through multiple specialists, enabling cross-domain reasoning and producing high-quality, verified outputs.\n\n### Key Features\n\n- **12 Domain Specialists** - Math, Physics, Chemistry, Biology, Code, Research, Memory, Facts, Simulation, Image, Sound, and General\n- **Web-of-Thought (WoT)** - Multi-step reasoning that chains specialists together dynamically\n- **Research Mode** - Generates academic-style PDF reports with proper citations\n- **Diary Memory** - Persistent memory across sessions for context continuity\n- **Self-Verification** - Built-in verifier ensures output quality and safety\n- **Runs Locally** - Uses quantized models (GGUF) via llama-cpp-python, no API keys required\n\n---\n\n## Quick Start\n\n### Prerequisites\n\n- Python 3.9 - 3.13 (3.13 recommended)\n- 8GB+ RAM (16GB recommended for Research Mode)\n- ~10GB disk space for models\n\n### Installation\n\n```bash\n# Clone the repository\ngit clone https://github.com/ra2157218-boop/Artificial-Neural-Mesh-V0.git\ncd Artificial-Neural-Mesh-V0\n\n# Create virtual environment\npython3 -m venv venv\nsource venv/bin/activate  # On Windows: venv\\Scripts\\activate\n\n# Install dependencies\npip install -r requirements.txt\n\n# Run ANM\npython run.py\n```\n\n### First Run\n\nOn first run, ANM will automatically download required models from HuggingFace (~3-5GB). This is a one-time process.\n\n```\nANM [Ready]\u003e What is quantum entanglement?\n```\n\n---\n\n## Architecture\n\nANM implements a **five-layer architecture** separating input, routing, specialist execution, output processing, and delivery.\n\n\u003cp align=\"center\"\u003e\n  \u003cimg src=\"docs/architecture.png\" alt=\"ANM Architecture\" width=\"700\"\u003e\n\u003c/p\u003e\n\n### Components\n\n| Layer | Component | Description |\n|-------|-----------|-------------|\n| **Layer 1** | Input, Classify, Memory | Query intake, classification, context retrieval |\n| **Layer 2** | Router, Planner, Metacog | Domain detection, execution planning, self-assessment |\n| **Layer 3** | 12 Specialists + WoT | Domain experts with Web-of-Thought orchestration |\n| **Layer 4** | Refiner, Verifier, Law Book | Output composition, quality validation, constitutional governance |\n| **Layer 5** | Response, PDF Gen, Session | Delivery formatting and session management |\n\n---\n\n## Research Mode\n\nResearch Mode provides in-depth analysis with academic-style PDF output.\n\n```bash\n# Enable Research Mode\npython run.py --research\n```\n\n### Features\n\n- **Academic PDF Output** - Two-column layout, proper sections\n- **Authority Models** - Domain-specific model assignments\n- **Web Search** - DuckDuckGo integration for current information\n- **Source Citations** - Tracks and cites all sources used\n- **Meta-Cognition** - Self-auditing of reasoning quality\n\n### PDF Sections\n\n1. Abstract \u0026 Keywords\n2. Introduction\n3. Methods (Authority Models, WoT Steps)\n4. Results (Domain Analysis)\n5. Discussion (Meta-Cognition)\n6. Limitations\n7. Sources\n8. Appendix (Technical Details)\n\n---\n\n## Models Used\n\nANM uses quantized GGUF models for efficient local inference:\n\n| Model | Size | Usage |\n|-------|------|-------|\n| DeepSeek-R1-Distill-Qwen-1.5B | ~1GB | General reasoning, routing |\n| Nanbeige4-3B | ~2GB | Math, Physics, Chemistry, Biology |\n| Stable-Code-3B | ~2GB | Code generation |\n| Qwen2.5-3B-Instruct | ~2GB | Internet research |\n\nModels are automatically downloaded from HuggingFace on first use.\n\n---\n\n## Benchmark Dataset\n\n\u003e **Official benchmark results and WoT traces are available on HuggingFace. This is the authoritative source for ANM performance metrics.**\n\n\u003cp align=\"center\"\u003e\n  \u003ca href=\"https://huggingface.co/datasets/Abd0r/anm-v0-benchmark\"\u003e\n    \u003cimg src=\"https://img.shields.io/badge/Dataset-Abd0r/anm--v0--benchmark-FFE066?style=for-the-badge\u0026logo=huggingface\" alt=\"HuggingFace Dataset\"\u003e\n  \u003c/a\u003e\n\u003c/p\u003e\n\nThe dataset includes:\n- **14 benchmark queries** across 9 domains (Math, Physics, Code, Chemistry, Biology, General, Cross-domain, Research, Memory)\n- **Complete WoT execution traces** showing how queries are routed through specialists\n- **Performance metrics** including latency, verification scores, and domain usage\n- **Structured query files** organized by domain for easy analysis\n\n### Load the Dataset\n\n```python\nfrom datasets import load_dataset\n\n# Load benchmark results\ndataset = load_dataset(\"Abd0r/anm-v0-benchmark\")\n\n# Or download specific files\nfrom huggingface_hub import hf_hub_download\nimport json\n\nmath_queries = json.load(open(hf_hub_download(\n    repo_id=\"Abd0r/anm-v0-benchmark\",\n    filename=\"queries/math.json\",\n    repo_type=\"dataset\"\n)))\n```\n\n---\n\n## Configuration\n\n### Environment Variables\n\n```bash\n# Memory settings\nexport ANM_MEMORY_ENABLED=true\nexport ANM_DIARY_FILE=anm_diary.txt\n\n# Research Mode\nexport ANM_RESEARCH_MODE=true\nexport ANM_PDF_OUTPUT=true\n```\n\n### Research Mode Config\n\nLocated in `anm/config/settings.py`:\n\n```python\nRESEARCH_MODE_CONFIG = {\n    \"authority_models\": {...},      # Model assignments per domain\n    \"workers_per_module\": 1,        # Deterministic (no ensemble)\n    \"wot_min_depth\": 3,             # Minimum reasoning steps\n    \"pdf_output\": True,             # Generate PDF\n}\n```\n\n---\n\n## Project Structure\n\n```\nANM-V0-OpenSource/\n├── anm/\n│   ├── __init__.py          # Main ANM interface\n│   ├── config/              # Configuration settings\n│   ├── router/              # Query routing \u0026 WoT orchestration\n│   ├── specialists/         # 12 domain specialists\n│   ├── wot/                 # Web-of-Thought engine\n│   ├── memory/              # Diary, working memory, learning\n│   ├── refiner/             # Output composition\n│   ├── verifier/            # Quality verification\n│   ├── output/              # PDF/Markdown generators\n│   └── system/              # Inference, model management\n├── run.py                   # Main entrypoint\n├── requirements.txt         # Dependencies\n└── README.md\n```\n\n---\n\n## Development\n\n### Tools Used\n\nThis project was developed with the assistance of:\n\n- **[Cursor](https://cursor.sh)** - AI-powered code editor\n- **[GPT-5.1,GPT-5.2](https://openai.com)** - OpenAI's language model\n- **[Claude Code](https://claude.ai)** - Anthropic's Claude for code assistance\n\n### Contributing\n\n1. Fork the repository\n2. Create your feature branch (`git checkout -b feature/AmazingFeature`)\n3. Commit your changes (`git commit -m 'Add some AmazingFeature'`)\n4. Push to the branch (`git push origin feature/AmazingFeature`)\n5. Open a Pull Request\n\n---\n\n## Troubleshooting\n\n### Common Issues\n\n**llama-cpp-python installation fails:**\n```bash\npip install llama-cpp-python --no-cache-dir\n```\n\n**CUDA/Metal support:**\n```bash\n# For CUDA (NVIDIA)\nCMAKE_ARGS=\"-DLLAMA_CUBLAS=on\" pip install llama-cpp-python\n\n# For Metal (Apple Silicon)\nCMAKE_ARGS=\"-DLLAMA_METAL=on\" pip install llama-cpp-python\n```\n\n**Models not downloading:**\n- Check internet connection\n- Ensure HuggingFace Hub is accessible\n- Try: `huggingface-cli login`\n\n**PDF generation fails:**\n- Install reportlab: `pip install reportlab`\n- Ensure write permissions in output directory\n\n---\n\n## License\n\nThis project is licensed under the MIT License - see the [LICENSE](LICENSE) file for details.\n\n---\n\n## Acknowledgments\n\n- [DeepSeek](https://github.com/deepseek-ai) for R1 reasoning models\n- [Qwen](https://github.com/QwenLM) for instruction-tuned models\n- [llama.cpp](https://github.com/ggerganov/llama.cpp) for efficient inference\n- [HuggingFace](https://huggingface.co) for model hosting\n\n---\n\n\u003cp align=\"center\"\u003e\n  \u003cb\u003eBuilt with AI, for AI reasoning\u003c/b\u003e\n  \u003cbr\u003e\n  \u003csub\u003eANM V0-OpenSource - Multi-Agent Reasoning System\u003c/sub\u003e\n\u003c/p\u003e\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fabd0r%2Fartificial-neural-mesh-v0","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fabd0r%2Fartificial-neural-mesh-v0","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fabd0r%2Fartificial-neural-mesh-v0/lists"}