Key Capabilities for Technical IT Users
Seamless Integrations
Connect MyAi with external systems, databases, and APIs to create unified data flows and automated processes. Leverage credential management for secure access.
Custom Functions & APIs
Develop, test, and deploy custom Python functions within MyAi to handle complex logic, data transformations, or specific business rules.
Data Orchestration
Design and manage how data moves through MyAi, ensuring data integrity, security, and performance across various artifacts and dimensions.
System Architecture
Understand how MyAi scales and how to architect solutions that are performant and maintainable for long-term use.
Relevant Tools & Concepts
API Client (`api_client`)
API Client (`api_client`)
For making HTTP API calls to external services. You can configure credentials and define complex requests directly from workflows or custom functions.
Custom Logic (`functions`)
Custom Logic (`functions`)
Write and deploy custom Python code that orchestrates tool calls, processes data, and implements business logic outside the standard built-in tools.
Database Access (`sql_client`)
Database Access (`sql_client`)
Execute SQL queries against various databases (BigQuery, MySQL, PostgreSQL), enabling direct data interaction and orchestration.
Integration Credentials
Integration Credentials
Securely store and manage authentication details for external systems, ensuring that your integrations are both functional and compliant.
Advanced Architecture & Integrations
For Technical IT Users, securing your integrations and custom code within MyAi is paramount. MyAi provides robust mechanisms to ensure your solutions are not only functional but also secure and compliant.Integration Credentials
These artifacts are the secure vault for all your external system authentication details (API keys, OAuth tokens, username/passwords).Best Practice: Never hardcode credentials in your
functions. Access to credential artifacts is managed via MyAi’s strict permission system.Functions Runtime Security
Custom Python
functions run within a secure, isolated environment.Principle of Least Privilege: When designing functions, ensure they only have access to the specific tools and data necessary for their operation. All executions are audited.Webhooks as Event Listeners
MyAi
workflows can be configured with webhook triggers. An external system (like a CRM, ERP, IoT device) can send an HTTP POST request to initiate a workflow.Use Cases: Real-time data synchronization, event-driven automation.API Client for Outbound
Conversely, MyAi
functions and workflows can use the api_client tool to call out to external APIs. This allows MyAi to push data, retrieve info, or trigger actions in other systems.Dimension Navigation (transition_dimension)
For Technical IT Users building complex workflows or integrations, understanding how to navigate and leverage different dimensions is crucial. The transition_dimension tool allows your custom logic to programmatically switch the AI’s current context.
When the AI transitions to a new dimension, it gains access to that dimension’s specific context, tools, and expertise. This is essential for multi-dimensional workflows where different stages of a process require specialized knowledge.
- Multi-Stage Workflows: Design workflows that initiate in one dimension (e.g., “Data Ingestion”) and then hand off to another specialized dimension (e.g., “Compliance Review”) for subsequent steps.
- Specialized Processing: Route tasks to the dimension best equipped to handle them. For example, a workflow might
transition_dimensionto a “Security Audit” dimension when a vulnerability is detected.
transition_dimension to a “Compliance Check” dimension for KYC verification, and finally to a “CRM Update” dimension.
Getting Started as an IT User
Review the architecture
Understand how Dimensions, Skills, Artifacts, and Workflows connect. This is the foundation for everything you’ll build.
Set up Integration Credentials
Before connecting external systems, create secure credential Artifacts for your API keys, OAuth tokens, and database connections. Never hardcode credentials in Functions.
Build your first Function
Write a Python-based Function that uses
api_client or sql_client to pull data from an external system. Test it in isolation before wiring it into a Workflow.Create a Workflow
Chain your Function with other steps — data transformation, Canvas generation, or notifications. Set a trigger (manual, scheduled, or webhook) and validate end-to-end using Work Orders.
Assign Skills to Dimensions
Bundle your Functions, Canvases, and Templates into Skills and map them to the appropriate Dimensions so end users can access the capabilities you’ve built.