🧱 Architecture Overview
Twazi is designed as a modular, extensible workflow automation platform. Its architecture enables seamless integration of AI, web automation, and blockchain technologies. Below is a high-level overview of the main components and how they interact.
Main Components
1. Workflow Engine
- Orchestrates the execution of workflows, managing the flow between tasks.
- Handles input/output passing, error handling, and execution state.
2. Task Library
- Collection of reusable tasks (e.g., AI reasoning, smart contract interaction, web scraping).
- Each task defines its own inputs, outputs, and logic.
- Extensible: users can add custom tasks.
3. Workflow Editor (UI)
- Visual interface for building and editing workflows.
- Drag, drop, and connect tasks visually.
- Provides real-time validation and feedback.
4. Credential & Security Manager
- Securely stores and manages API keys, private keys, and sensitive data.
- Ensures credentials are only accessible to authorized tasks.
5. Execution Monitor
- Tracks workflow runs, logs, and results.
- Displays step-by-step execution with inputs, outputs, and errors.
6. Credit System
- Manages billing and resource usage.
- Each task consumes credits based on complexity or API usage.
7. API & Integration Layer
- Provides endpoints for workflow management and execution.
- Connects Twazi to external AI, web, or blockchain services.
High-Level Interaction Diagram
graph TD
A[User] -->|Builds| B(Workflow Editor)
B -->|Saves| C(Workflow Engine)
C -->|Executes| D(Task Library)
D -->|Requests| E(Credential Manager)
D -->|Consumes| F(Credit System)
D -->|Logs| G(Execution Monitor)
C -->|APIs| H(API & Integration Layer)
H -->|External Services| I[AI, Blockchain, Web]
style A fill:#fef3c7,stroke:#000,stroke-width:2
style B fill:#fef3c7,stroke:#000,stroke-width:2
style C fill:#86efac,stroke:#000,stroke-width:2
style D fill:#86efac,stroke:#000,stroke-width:2
style E fill:#a5b4fc,stroke:#000,stroke-width:2
style F fill:#a5b4fc,stroke:#000,stroke-width:2
style G fill:#a5b4fc,stroke:#000,stroke-width:2
style H fill:#86efac,stroke:#000,stroke-width:2
style I fill:#a5b4fc,stroke:#000,stroke-width:2
1. Users design workflows in the visual editor.
2. Workflows are saved and managed by the engine.
3. Tasks run in sequence or parallel, passing data as needed.
4. Tasks interact with external services using credentials.
5. Execution Monitor logs and provides real-time feedback.
6. Credit System tracks and limits usage.
2. Workflows are saved and managed by the engine.
3. Tasks run in sequence or parallel, passing data as needed.
4. Tasks interact with external services using credentials.
5. Execution Monitor logs and provides real-time feedback.
6. Credit System tracks and limits usage.
🧩 This architecture makes Twazi flexible, secure, and extensible for any automation use case—from AI agents to DeFi bots.