Skip to content

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

160 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Supreme Architect of The Code Eternal. An advanced autonomous AI agent system with cognitive architecture (Orythix), local multi-model routing, self-healing incident response, and high-performance terminal interface.

VIKI: Sovereign Autonomous AI Agent & Local LLM Engineer

Local LLM Orchestration • Air-Gapped Privacy • Autonomous Self-Evolution • Full-SDLC Engineering Swarm

Build & Tests Mypy Strict Ruff Python 3.10+ Local LLMs Cloud LLMs License: Apache 2.0 Docker


Key FeaturesChinese Overview / 中文说明Multi-Model Provider MatrixArchitectureQuick StartEngineering PlaybooksNeural ForgeFAQLicense


What is VIKI? (Project Overview)

VIKI (Virtual Intelligence Knowledge Interface) is a production-grade, self-hosted autonomous AI agent and developer-first AI engineer written in Python. Built for local-first privacy and zero data leakage, VIKI runs natively on your hardware via LM Studio and Ollama, while seamlessly supporting cloud frontier models (OpenAI GPT-4o, Anthropic Claude 3.7 Sonnet, Google Gemini 2.5 Pro).

Powered by the Orythix Cognitive Architecture, VIKI provides deep ReAct reasoning, multi-agent swarm orchestration, automated incident healing (Sentry/Datadog stack trace ingestion), persistent SQLite RAG memory, and self-evolution through the Neural Forge—all accessible via an ultra-fast CLI-first REPL and web dashboard.


Chinese Overview / 中文说明

VIKI(虚拟智能知识接口)是一个高性自托管自主 AI 智能体与开发者优先的 AI 工程师系统。VIKI 专为本地优先与隐私安全设计,能够在无数据泄漏的前提下在本地硬件上运行(基于 LM StudioOllama),同时也无缝支持云端顶级大模型(OpenAI GPT-4oAnthropic Claude 3.7 SonnetGoogle Gemini 2.5 Pro)。

核心优势 (Key Highlights in Chinese):

  • 100% 本地与物理隔离: 默认操作无需依赖外网 API(开启 VIKI_AIR_GAP=1 模式),保障核心代码与数据安全。
  • Orythix 认知架构: 具备五层认知大脑(感知、判断、思考、反思、执行与元认知),在执行前自动审计计划,有效降低幻觉。
  • 多模型灵活路由: 自动分发任务至本地模型(LM Studio/Ollama)或云端模型(OpenAI/Claude/Gemini)。
  • 全生命周期软件工程与 Jira 自动化: 自动解析 Jira 需求、生成技术规范、编写 TDD 测试套件、执行 AST 代码重构以及生成 OpenAPI/gRPC 接口。
  • 自主故障修复: 支持对接 Sentry/Datadog 告警,在隔离的 Git Worktree 中重现 Bug 并自动修复。
  • 神经铸造厂 (Neural Forge): 无需 GPU 训练,将交互经验与 SQLite 记忆直接“烘焙”到本地模型的 System Prompt 中。

Key Features

  • 100% Sovereign & Air-Gap Capable: Default operation never leaves your local network (VIKI_AIR_GAP=1). Zero telemetry, zero external tracking.
  • Orythix 5-Layer Cognitive Cortex: Judgment, Deliberation, Reflection, Execution, and Meta-Cognition layers reduce hallucinations and audit plans before execution.
  • Multi-Model Provider Routing: Instantly switch or auto-route between LM Studio, Ollama, OpenAI, Anthropic Claude, and Google Gemini.
  • Full-SDLC & Jira Automation: Automated Jira ticket parsing, technical spec generation, TDD test suites, AST codemods, and OpenAPI 3.1 / gRPC proto generation.
  • Autonomous Incident Healing: Connects to Sentry/Datadog webhooks, reproduces bugs in isolated Git worktrees, applies targeted fixes, verifies via pytest, and creates PRs.
  • Multi-Agent Swarm Engineering: Execute complex DAG tasks using leader-worker agent swarms and Monte Carlo Tree Search (MCTS) reasoning.
  • Local Neural Forge: Bakes user feedback and lessons accumulated in SQLite directly into custom model system prompts without GPU training.
  • Low-Hardware Optimization: Runs efficiently on 8 GB RAM laptops (with dedicated 4k token caps, prompt compression, and LRU entity extraction caching).
  • Capability Gating & Security: Sandboxed file and shell execution (viki/core/sandbox.py) with strict permission boundaries and prompt injection detection.

Multi-Model Provider Matrix

VIKI dynamically routes queries to the optimal provider profile configured in config/models.yaml:

Provider Supported Models / Endpoints Privacy / Mode Best For
LM Studio gemma-4-e4b, qwen3.5-9b, deepseek-r1 (http://localhost:1234/v1) 100% Local / Air-Gapped Default offline coding & privacy
Ollama llama3, codellama, mistral, phi3 (http://localhost:11434/v1) 100% Local / Air-Gapped Local CLI automation & zero-cost tasks
OpenAI gpt-4o, gpt-4o-mini, o3-mini Cloud API Deep architectural reasoning & code review
Anthropic claude-3-7-sonnet, claude-3-5-haiku Cloud API Full-SDLC engineering & complex refactoring
Google gemini-2-5-pro, gemini-2-5-flash Cloud API Multimodal vision & massive context research

Technical Architecture

VIKI executes tasks through a deterministic 5-Layer Consciousness Cortex:

graph TD
    A["User Prompt / Event Trigger"] --> B["Layer 1: Perception & Request Pipeline"]
    B --> C["Layer 2: Judgment Engine (Reflex vs Deliberation)"]
    C -->|Reflex Path| D["Instant Short-Circuit Execution"]
    C -->|Deliberate Path| E["Layer 3: Cortex & Multi-Model Router"]
    E --> F["Layer 4: Reflection & Hallucination Audit"]
    F --> G["Layer 5: Execution & Capability Registry"]
    G --> H["Layer 6: Meta-Cognition & Memory Persistence"]
    H --> I["SQLite RAG Wisdom DB & Neural Forge"]
Loading

Quick Start

Prerequisites

  • Python 3.10+ (3.10, 3.11, and 3.12 supported)
  • Local Inference Server: LM Studio or Ollama running locally (http://localhost:1234 or http://localhost:11434)

1. Installation

# Clone repository
git clone https://github.com/Orythix/viki.git
cd viki

# Create & activate virtual environment
python -m venv .venv
# On Windows (PowerShell):
./.venv/Scripts/Activate.ps1
# On Linux/macOS:
source .venv/bin/activate

# Install core package
pip install -e .

2. Configure Environment

cp .env.example .env

(Optionally set OPENAI_API_KEY, ANTHROPIC_API_KEY, GEMINI_API_KEY, or custom VIKI_PERSONA in .env)

3. Launch VIKI

# Start interactive CLI REPL in current workspace
viki

# Single-shot command (runs task and exits)
viki "refactor auth logic in src/viki/core/security.py"

# Run with Dev Persona (coding-focused skills)
VIKI_PERSONA=dev viki

Docker Deployment

Run VIKI in an isolated container using Docker Compose:

# Build and run interactive container
docker compose build
docker compose run --rm -it viki

See docs/DOCKER.md for advanced container configurations.


Engineering Playbooks & SDLC

VIKI includes 40+ built-in engineering playbooks spanning the complete software development lifecycle:

  • Define & Plan: idea_refine, spec_driven_development, planning_and_task_breakdown, jira_sdlc_workflow
  • Build & Refactor: incremental_implementation, test_driven_development, ast_codemod, openapi_schema, frontend_ui_engineering
  • Verify & Review: browser_testing_with_devtools, debugging_and_error_recovery, code_review_and_quality, security_and_hardening
  • Ship & Recover: git_workflow_and_versioning, ci_cd_and_automation, autonomous_incident_healing

Build Your VIKI Model (Neural Forge)

Bake your accumulated SQLite lessons and project wisdom directly into custom LM Studio model profiles without GPU training:

# Bake top lessons into data/Modelfile.viki_evolved
python scripts/build_viki_model.py

# Set models.yaml default profile to lmstudio-gemma4e4b
python scripts/build_viki_model.py --set-default

Comparison: VIKI vs Other AI Agents

Feature VIKI AutoGPT LangChain Manus / Devin
Air-Gapped Privacy Yes (100% Local) No (Cloud API) Limited (Library only) No (Cloud Hosted)
Local LLM Native Yes (LM Studio / Ollama) Limited Requires custom code No (Cloud Hosted)
Incident Healing (Sentry) Yes (Built-in) No No Proprietary
Neural Forge Evolution Yes (Built-in) No No No
MCTS Swarm Execution Yes (Built-in) No Manual DAG required Proprietary
Open Source License Yes (Apache-2.0) Yes (MIT) Yes (MIT) Closed Source

Frequently Asked Questions

Is VIKI completely free and open source? Yes. VIKI is released under the Apache-2.0 open source license. You can self-host, modify, and deploy VIKI freely without any required cloud subscription.
Does VIKI support cloud AI models like Claude or GPT-4o? Yes. VIKI features a unified provider adapter supporting LM Studio, Ollama, OpenAI (GPT-4o), Anthropic (Claude 3.7 Sonnet), and Google Gemini (2.5 Pro). Configure your API keys in .env or config/models.yaml.
Can VIKI run on laptops with 8 GB RAM? Yes. VIKI is specifically optimized for low-spec hardware (4 GB / 8 GB RAM) with aggressive token capping, prompt compression, LRU entity extraction caching, and lazy skill loading.

Documentation Index


Keywords & Topics

Local AI AgentAutonomous AI EngineerSelf-Hosted AI AssistantLM Studio AgentOllama AgentAir-Gapped AISovereign AIReAct ReasoningModel Context Protocol (MCP)Agentic SDLCSentry Bug HealingNeural ForgePrivate LLMPython AI Agent中文 AI 智能体本地大模型


License

VIKI is licensed under the Apache License 2.0.

About

Autonomous AI Engineer & Sovereign Local Agent. Air-gapped privacy, Orythix cognitive cortex, automated Sentry bug healing, and Neural Forge prompt baking. Runs 100% locally on LM Studio & Ollama + Cloud (Claude 3.7 / GPT-4o). ⚡ 487/487 Tests Passing.

Topics

Resources

Code of conduct

Contributing

Security policy

Stars

2 stars

Watchers

1 watching

Forks

Used by

Contributors

Languages