OriCore

A powerful, standalone AI engine with multi-model support, tool calling, and extensible architecture

English | 中文文档

About OriCore

OriCore is a comprehensive AI engine that provides the core functionality for building intelligent assistants. It offers a rich set of tools, multi-model support, and extensible architecture through MCP (Model Context Protocol) and Skills.

Key Features

• Multi-Model Support: Compatible with 40+ AI providers including OpenAI, Anthropic, Google, DeepSeek, and more
• Rich Tool System: Built-in tools for file operations (read, write, edit), code search (grep, glob), shell commands, and web requests
• Interaction Modes: Specialized modes for different tasks - brainstorming, planning, code review, debugging, and default
• MCP Integration: Extensible via Model Context Protocol for custom tools and capabilities
• Skill System: Load and use custom skills from local or remote sources (GitHub, GitLab)
• Agent Framework: Built-in agents for complex multi-step tasks (explore, general-purpose)
• Session Management: Persistent conversation history with session resumption
• Streaming Support: Real-time text delta streaming for responsive interactions
• Fully Typed: Complete TypeScript support with comprehensive type definitions
• Zero Configuration: Works out of the box with sensible defaults

Use Cases

• Build custom AI assistants
• Integrate AI capabilities into IDE extensions
• Create automated code review systems
• Develop intelligent debugging tools
• Build educational platforms
• Implement automated documentation generation

Installation

npm install oricore

# Using pnpm
pnpm add oricore

# Using bun
bun add oricore

Optional Dependencies

OriCore has support for additional features through optional dependencies:

PDF Support

npm install pdf-parse

The read tool can parse PDF files when pdf-parse is installed. Without it, PDF reading will be disabled.

Quick Start

import { createEngine } from 'oricore';

// 1. Create the engine
const engine = createEngine({
  productName: 'MyAIAssistant',
  version: '1.0.0',
});

// 2. Initialize with model and API key
await engine.initialize({
  model: 'openai/gpt-5.2-codex',
  provider: {
    openai: {
      apiKey: 'your-api-key',
      baseURL: 'https://api.openai.com/v1',
    },
  },
});

// 3. Send a message
const result = await engine.sendMessage({
  message: 'Create a TypeScript function to calculate fibonacci',
  write: true,
});

console.log(result.data.text);

// 4. Cleanup
await engine.shutdown();

Interaction Modes

OriCore provides specialized modes for different tasks:

// Brainstorm mode - interactive design exploration
engine.setMode('brainstorm');
const design = await engine.sendMessageWithMode('I want to build a task management app');

// Plan mode - create implementation plans
engine.setMode('plan');
const plan = await engine.sendMessageWithMode('Create a plan for adding user authentication');

// Review mode - code review and analysis
engine.setMode('review');
const review = await engine.sendMessageWithMode('Review this code: ...');

// Debug mode - troubleshooting
engine.setMode('debug');
const fix = await engine.sendMessageWithMode('Help debug this error: ...');

Built-in Tools

OriCore includes a comprehensive set of tools:

Tool	Description
read	Read file contents (supports text, images, and PDF*)
write	Write new files
edit	Edit existing files with search/replace
glob	Find files by pattern
grep	Search file contents
bash	Execute shell commands
fetch	Make HTTP requests
askUserQuestion	Interactive Q&A with users
task	Spawn specialized agents
todo	Track task progress

*PDF support requires the optional pdf-parse package (see below)

Configuration

Full Configuration Example

await engine.initialize({
  model: 'openai/gpt-5.2-codex',
  planModel: 'openai/gpt-5.2-codex',
  approvalMode: 'autoEdit',
  language: 'en',
  tools: {
    read: true,
    write: true,
    bash: true,
  },
  provider: {
    openai: {
      apiKey: 'your-api-key',
      baseURL: 'https://api.openai.com/v1',
    },
  },
});

Supported Providers

Provider	Model Example	API Base URL
OpenAI	openai/gpt-5.2-codex	https://api.openai.com/v1
Anthropic	anthropic/claude-opus-4-5	https://api.anthropic.com
Google	google/gemini-3-flash-preview	https://generativelanguage.googleapis.com
DeepSeek	deepseek/deepseek-chat	https://api.deepseek.com
Zhipu AI	zhipuai/glm-4.7	https://open.bigmodel.cn/api/paas/v4

See USAGE.md for more configuration options.

Project Structure

oricore/
├── src/
│   ├── api/           # Main Engine API
│   ├── core/          # Core functionality (loop, context, config)
│   ├── tools/         # Built-in tools
│   ├── modes/         # Interaction modes
│   ├── mcp/           # MCP integration
│   ├── skill/         # Skill system
│   ├── agent/         # Agent framework
│   ├── session/       # Session management
│   └── utils/         # Utilities
├── examples/          # Usage examples
└── dist/              # Compiled output

Statement

This project references the core architecture of the following excellent project:

• neovate-code - Core AI engine architecture

OriCore has been refactored and streamlined on this foundation, removing UI, CLI, and other peripheral features to focus on providing a lightweight, standalone AI engine library that can be easily integrated into any project.

License

MIT © lyw405

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

Support

If you have any questions or issues, please open an issue on GitHub.