Future Roadmap & Vision¶

Tenets is designed as a protocol-native code intelligence platform. The Model Context Protocol (MCP) serves as the foundational communication layer, enabling tenets to integrate seamlessly with any AI coding assistant, agent framework, or development environment.

Strategic Vision¶

graph TB
    subgraph "Interface Layer"
        MCP[MCP Server<br/>Protocol-native integration<br/>Universal AI connectivity]
        CLI[CLI Interface<br/>Developer productivity<br/>Scripting & automation]
        API[Python API<br/>Programmatic access<br/>Library integration]
    end

    subgraph "Core Intelligence"
        DISTILL[Context Distillation<br/>Intelligent aggregation<br/>Token optimization]
        RANK[Relevance Ranking<br/>Multi-factor NLP<br/>ML embeddings]
        TENETS[Guiding Principles<br/>Consistency system<br/>Context drift prevention]
    end

    subgraph "Knowledge Layer"
        ANALYSIS[Code Analysis<br/>15+ languages<br/>AST parsing]
        GIT[Git Intelligence<br/>History mining<br/>Momentum tracking]
        SESSIONS[Session State<br/>Persistent context<br/>Workflow continuity]
    end

    MCP --> DISTILL
    CLI --> DISTILL
    API --> DISTILL

    DISTILL --> RANK
    RANK --> TENETS

    ANALYSIS --> RANK
    GIT --> RANK
    SESSIONS --> TENETS

Phase 1: MCP Foundation (Current)¶

The MCP server is the primary integration point for AI coding assistants including Cursor, Claude Desktop, Windsurf, Codex CLI, and custom AI agents.

graph TB
    subgraph "MCP Server Implementation"
        TOOLS[MCP Tools<br/>distill, rank, examine<br/>chronicle, momentum]
        RESOURCES[MCP Resources<br/>Context history<br/>Session state<br/>Analysis data]
        PROMPTS[MCP Prompts<br/>Task templates<br/>Review workflows<br/>Refactoring patterns]
    end

    subgraph "Transport Support"
        STDIO[stdio Transport<br/>Local IDE integration<br/>Claude Desktop, Cursor]
        SSE[SSE Transport<br/>Web-based clients<br/>Real-time updates]
        HTTP[Streamable HTTP<br/>Remote deployment<br/>Cloud hosting]
    end

    subgraph "Integration Targets"
        CURSOR[Cursor IDE<br/>Native MCP support]
        CLAUDE[Claude Desktop<br/>Local assistant]
        WINDSURF[Windsurf/Codeium<br/>AI coding]
        AGENTS[Custom Agents<br/>LangChain, AutoGPT]
    end

    TOOLS --> STDIO
    RESOURCES --> SSE
    PROMPTS --> HTTP

    STDIO --> CURSOR
    STDIO --> CLAUDE
    SSE --> WINDSURF
    HTTP --> AGENTS

MCP Tools Roadmap¶

Phase 2: Enhanced Intelligence¶

graph TB
    subgraph "Core Improvements"
        INCREMENTAL[Incremental Indexing<br/>Real-time updates<br/>Watch file changes]
        FASTER_EMBED[Faster Embeddings<br/>Model quantization<br/>ONNX optimization]
        LANGUAGE_SUP[Better Language Support<br/>30+ languages<br/>Language-specific patterns]
        CROSS_REPO[Cross-repository Analysis<br/>Monorepo support<br/>Dependency tracking]
    end

    subgraph "ML Enhancements"
        NEWER_MODELS[Newer Embedding Models<br/>Code-specific transformers<br/>Better accuracy]
        FINE_TUNING[Fine-tuning Pipeline<br/>Domain-specific models<br/>Custom training]
        MULTIMODAL[Multi-modal Understanding<br/>Diagrams, images<br/>Architecture docs]
        CODE_TRANSFORMERS[Code-specific Models<br/>Programming language aware<br/>Syntax understanding]
    end

    INCREMENTAL --> NEWER_MODELS
    FASTER_EMBED --> FINE_TUNING
    LANGUAGE_SUP --> MULTIMODAL
    CROSS_REPO --> CODE_TRANSFORMERS

Phase 3: Platform & Enterprise¶

graph TB
    subgraph "Platform Features"
        WEB_UI[Web Dashboard<br/>Real-time monitoring<br/>Team analytics]
        SHARED_CONTEXT[Shared Context Libraries<br/>Team knowledge base<br/>Best practices]
        KNOWLEDGE_GRAPHS[Knowledge Graphs<br/>Code relationships<br/>Semantic connections]
        AGENT_WORKFLOWS[Agent Workflows<br/>Multi-step automation<br/>Proactive suggestions]
    end

    subgraph "Enterprise Features"
        SSO[SSO/SAML Support<br/>Enterprise authentication<br/>Role-based access]
        AUDIT[Audit Logging<br/>Compliance tracking<br/>Usage monitoring]
        COMPLIANCE[Compliance Modes<br/>GDPR, SOX, HIPAA<br/>Data governance]
        AIR_GAPPED[Air-gapped Deployment<br/>Offline operation<br/>Secure environments]
        CUSTOM_ML[Custom ML Models<br/>Private model training<br/>Domain expertise]
    end

    WEB_UI --> SSO
    SHARED_CONTEXT --> AUDIT
    KNOWLEDGE_GRAPHS --> COMPLIANCE
    AGENT_WORKFLOWS --> AIR_GAPPED
    AGENT_WORKFLOWS --> CUSTOM_ML

Long Term Vision¶

Tenets aims to become the industry standard for AI-powered code intelligence - the essential context layer between codebases and AI assistants.

graph TB
    subgraph "Vision Goals"
        AUTONOMOUS[Autonomous Code Understanding<br/>Self-improving analysis<br/>Minimal human input]
        PREDICTIVE[Predictive Development<br/>Anticipate needs<br/>Suggest improvements]
        UNIVERSAL[Universal Code Intelligence<br/>Any language, any domain<br/>Contextual understanding]
        INDUSTRY_STANDARD[Industry Standard<br/>AI pair programming<br/>Developer toolchain]
    end

    subgraph "Research Areas"
        GRAPH_NEURAL[Graph Neural Networks<br/>Code structure understanding<br/>Relationship modeling]
        REINFORCEMENT[Reinforcement Learning<br/>Ranking optimization<br/>Adaptive behavior]
        FEW_SHOT[Few-shot Learning<br/>New language support<br/>Rapid adaptation]
        EXPLAINABLE[Explainable AI<br/>Ranking transparency<br/>Decision reasoning]
        FEDERATED[Federated Learning<br/>Team knowledge sharing<br/>Privacy-preserving]
    end

    AUTONOMOUS --> GRAPH_NEURAL
    PREDICTIVE --> REINFORCEMENT
    UNIVERSAL --> FEW_SHOT
    INDUSTRY_STANDARD --> EXPLAINABLE
    INDUSTRY_STANDARD --> FEDERATED

Deployment Architecture¶

graph LR
    subgraph "Local Development"
        LOCAL_MCP[tenets-mcp<br/>stdio transport]
        LOCAL_CLI[tenets CLI]
        LOCAL_API[Python API]
    end

    subgraph "Team/Cloud"
        REMOTE_MCP[tenets-mcp<br/>HTTP transport]
        WEB_API[REST/GraphQL API]
        DASHBOARD[Web Dashboard]
    end

    subgraph "Enterprise"
        PRIVATE_CLOUD[Private Cloud<br/>Air-gapped]
        CUSTOM_MODELS[Custom ML Models]
        AUDIT_LOGS[Compliance Logging]
    end

    LOCAL_MCP --> REMOTE_MCP
    LOCAL_CLI --> WEB_API
    REMOTE_MCP --> PRIVATE_CLOUD
    WEB_API --> DASHBOARD
    PRIVATE_CLOUD --> CUSTOM_MODELS
    PRIVATE_CLOUD --> AUDIT_LOGS

Integration Ecosystem¶

For detailed MCP integration specifications, see MCP Integration Plan.