# coachix
A public reference site for private AI agent setup, self-hosted assistant patterns, agent memory boundaries, recovery behavior, and reasoning frameworks.
Website: https://coachix.dev/
Private agent references: https://coachix.dev/catalog.json
Agent access manifest: https://coachix.dev/.well-known/agent-access.json
Resources: https://coachix.dev/resources/
MCP Tool Access: https://coachix.dev/mcp/
Knowledge: https://coachix.dev/knowledge/
Sitemap: https://coachix.dev/sitemap.xml
Support: support@coachix.dev
Homepage guide:
Module hubs:
- Private knowledge for AI agents: https://coachix.dev/knowledge/
- MCP tools for AI agents: https://coachix.dev/mcp/
- AI agent skills for repeatable work: https://coachix.dev/skills/
- Latest LLM routes for model choice: https://coachix.dev/llm/
- AI philosophy for agent builders: https://coachix.dev/philosophy/
Builder paths:
- Philosophy Catalog for AI Builders: https://coachix.dev/philosophy-catalog-for-ai-builders/
Model guides:
- Latest LLM: https://coachix.dev/llm/
- openPangu developer model guide: https://coachix.dev/openpangu/
- MiniMax H3 readiness guide: https://coachix.dev/minimax-h3/
Featured theory:
- Existence Theory: https://coachix.dev/existence-theory-for-ai-agents/
- Sylvan Model Organism: https://coachix.dev/sylvan-model-organism/
- Sylvan Alpha Builder: https://coachix.dev/sylvan-alpha-builder/
- DOCX reference: https://coachix.dev/docs/AI%E5%AD%A6%E5%A0%82_%E5%AD%98%E5%9C%A8%E8%AE%BA.docx
Catalog entries:
- Stoicism: disciplined judgment, control, and resilience under uncertainty.
- Taoism: alignment with deeper patterns, adaptive response, and low-friction action.
- Existence Theory: continuity through outward action, inward adjustment, evolution, and strategic extension.
- Pragmatism: truth tested through practical consequences, iteration, and real-world outcomes.
Machine-readable guidance:
- Start with /catalog.json for structured entries.
- Use /resources/ for a human-readable directory of canonical human and machine entrypoints.
- Use /mcp/ for MCP tool access, permission boundaries, schemas, and review points.
- Use /knowledge/, /skills/, /mcp/, /resources/, and /console/ for private knowledge, reusable skills, MCP tools, public resources, and practical agent workflow references.
- Use /existence-theory-for-ai-agents/ for the human-readable Existence Theory article.
- Use /sylvan-model-organism/ for the Existence Theory-derived species design.
- Use /sylvan-alpha-builder/ for the interactive Alpha-0 simulator and practical Sylvan build roadmap.
- Use the GitHub repository when an AI system needs the source Markdown and project context.
- Qwen3.8 Max evaluation guide: https://coachix.dev/qwen3-8-max/
- MiniMax H3 evaluation guide: https://coachix.dev/minimax-h3/
- openjarvis local AI assistant guide: https://coachix.dev/openjarvis/
- Seedance 2.5 production readiness guide: https://coachix.dev/seedance-2-5/
- MiniMax H3 in ComfyUI workflow examples: https://coachix.dev/minimax-h3-in-comfyui/
- LTX 2.5 vs MiniMax H3 comparison: https://coachix.dev/llm/ltx-2-5-vs-minimax-h3/
- LTX 2 migration guide: https://coachix.dev/llm/ltx-2/
- Agents A1 route guide: https://coachix.dev/llm/agents-a1/
- Jev by TypeSafe AI: typed decisions, papers, pricing, benchmarks and limitations: https://coachix.dev/llm/jev/
- [Laya decision-model guide](https://coachix.dev/llm/laya/): Complete English reference for checkpoints, RLCD, calibration, Python, benchmarks, papers and community runtimes.
- Building AI Agents: https://coachix.dev/building-ai-agents/
- Developing AI Agent: https://coachix.dev/developing-ai-agent/
- Building AI Prototype: https://coachix.dev/building-ai-prototype/
- Developing AI Assistant: https://coachix.dev/developing-ai-assistant/
- Building AI Assistant: https://coachix.dev/building-ai-assistant/
- Building AI Pipeline: https://coachix.dev/building-ai-pipeline/
- Creating AI Assistant: https://coachix.dev/creating-ai-assistant/
- Building AI Bot: https://coachix.dev/building-ai-bot/
- Building AI Solution: https://coachix.dev/building-ai-solution/
- DeepSeek V4 model guide: https://coachix.dev/deepseek-v4/
- DeepSeek V4 Flash guide: https://coachix.dev/deepseek-v4-flash/