Search & Metadata

These tools let your custom code "see" the rest of the system. Use them to find artifacts, build dynamic workflows, and maintain the architectural connections between components.

---

search_artifacts — Programmatic Discovery

The primary way functions discover and retrieve artifacts within MyAi. Supports both natural language and structured queries.

Semantic Search

Uses vector embeddings to find artifacts by meaning:

results = default_api.search_artifacts(
    operation="semantic",
    query="Find all documents related to the Q3 Audit",
    limit=20
)

Structured Queries

For precise, programmatic lookups:

Query Type	Purpose	Key Parameters
find_instances	All records created from a specific Template	template_id
find_by_variable	Filter by a Template's global variable value	template_id, variable_name, variable_value
find_by_category	Schema discovery across a type	category (e.g., canvases, functions, workflows)
find_recent	Recently created or updated artifacts	limit
count	Count of matching artifacts	varies

Example: Find Open Support Tickets

def main(team_lead: str):
    results = default_api.search_artifacts(
        operation="query",
        query_type="find_by_variable",
        template_id="support-ticket-template-id",
        variable_name="status",
        variable_value="Open",
        limit=50
    )
    tickets = results.get("artifacts", [])
    for ticket in tickets:
        subject = ticket.get("global_variables", {}).get("subject")
        print(f"  Ticket: {ticket.get('artifact_id')} — {subject}")
    return tickets

---

relationships — The Context Graph

The relationships tool manages the connections between artifacts and Dimensions. It builds self-documenting systems where every component knows why it's connected to others.

Relationship Types

Type	Meaning	Example
delegates_to	Workflow trust	A Workflow delegates execution to a Function
governs	Schema/permission control	A Dimension governs a set of Templates
implements	Code-to-interface link	A Function implements a Skill's contract
related_to	General association	A Canvas relates to a Knowledge Base article
parent_of	Hierarchical containment	A parent artifact contains child components

Perspective Descriptions

Each relationship includes natural language metadata — from_perspective and to_perspective — that captures why two things are connected. This allows the AI to reason about system architecture.

default_api.relationships(
    operation="create",
    from_id="sales-dashboard-canvas",
    to_id="sales-dimension",
    relationship_type="governs",
    from_perspective="I display KPIs for this dimension's sales team.",
    to_perspective="I govern this dashboard and provide its data context."
)

Operations

Operation	Purpose
create	Establish a new relationship
update	Modify an existing relationship's descriptions
archive	Deactivate a relationship
get_network	Retrieve the graph around an entity or the entire system
get_function_dependencies	Audit exactly which tools and credentials a function uses

Tip: Use get_function_dependencies before deploying a function to production — it lets IT audit every tool and credential the function touches.

---

Learn More

• Functions — Write custom logic that queries artifacts and manages relationships.
• Architecture — The big picture of how Dimensions, Skills, and Artifacts connect.