repository·main·Indexed 27 days ago
https://github.com/bytedance/flowgram.aiA composable and extensible workflow development framework for building AI workflow platforms. It provides core building blocks such as flow canvases, node configuration forms, and variable management engines. The framework supports both fixed-layout and free-layout editors, featuring a modular plugin architecture, a registry pattern for custom node types, and full TypeScript support.
The fixed-layout demo serves as a best-practices implementation for fixed-layout workflow development. It utilizes the following core technologies:
@douyinfe/semi-ui)FlowGram is a composable, visual, and extensible workflow development framework designed to help developers build AI workflow platforms. Key features include:
FlowGram Runtime is a reference implementation of a workflow runtime engine designed for business developers to learn from and modify. It parses and executes graph-based workflows using various node types.
⚠️ Critical Limitations:
FlowGram is a composable, visual, and extensible workflow development framework and toolkit designed to help developers build AI workflow platforms. Key features include:
FlowGram is a specialized workflow development framework and toolkit designed to help developers build AI workflow platforms. Unlike general-purpose graphics libraries, FlowGram provides built-in tools specifically for workflow scenarios, reducing the complexity of managing node data, dynamic forms, and variable scopes.
Core features include:
The FlowGram variable engine uses several core concepts to manage data during design and runtime:
VariableDeclaration (globally unique) and Property (unique within an object).StringType, IntegerType, ObjectType, ArrayType, MapType, CustomType).FlowGram is a composable, visual, and extensible workflow development framework designed for building AI workflow platforms. Key features include:
The workflow engine is the core component responsible for parsing, scheduling, and executing workflows. It is primarily implemented in js-core/src/domain/engine/engine.ts.
Key responsibilities include:
// Core method for workflow execution
public async run(params: RunParams): Promise<RunResult> {
const { schema, inputs, options } = params;
// Create workflow context
const context = this.createContext(schema, inputs, options);
try {
// Initialize workflow
await this.initialize(context);
// Execute workflow
await this.execute(context);
// Get workflow result
const result = await this.getResult(context);
return {
status: 'success',
outputs: result
};
} catch (error) {
// Error handling
return {
status: 'fail',
error: error.message
};
}
}FlowGram is designed as a complete out-of-the-box solution for flow editing, whereas ReactFlow focuses on a low-level rendering engine (Nodes, Edges, Handles) that requires significant upper-layer development for complex scenarios like fixed layouts.
Many features that are paid in ReactFlow are supported natively in FlowGram, including:
Note: ForceLayout is currently not supported by FlowGram, and Collaborative features are planned for the future.
The FlowGram.AI Free Layout demo is built on a highly modular and extensible architecture designed for workflow development. Key architectural features include:
FlowGram.AI is built on several core principles designed for workflow development:
The Free Layout demo is built upon the @flowgram.ai/free-layout-editor framework, which provides the foundational capabilities for free-form canvas editing. It includes: