Agent Types in the Gallery

Use this quick reference to understand the primary agent categories available in the gallery and how they complement each other during solution design.

Orchestrator Agents

Orchestrator agents coordinate multiple specialized agents or services to deliver an end-to-end workflow. They handle routing, state management, and error recovery so domain experts can focus on their tasks. Reach for an orchestrator when:

You need to compose several utility agents into a cohesive experience.
Work needs centralized observability or policy enforcement.
Stakeholders expect one entry point that adapts to user intent in real time.

Super Agents

Super agents encapsulate broad, high-impact capabilities that solve complex problems without additional orchestration. They typically blend reasoning, tool use, and memory within a single agent. Super agents are ideal when:

Users require a generalist agent that can triage diverse requests.
You want to accelerate adoption with fewer integration dependencies.
The offering needs to showcase cutting-edge functionality quickly.

Utility Agents

Utility agents specialize in a narrow task such as data enrichment, translation, transcription, or document retrieval. They operate as reusable building blocks that other agents can call on demand. Consider a utility agent when:

A workflow needs a reliable, auditable micro-service powered by AI.
You want to replace or augment deterministic automation with generative logic.
Teams require consistent outputs that can be validated easily.

MCP Agents

Model Context Protocol (MCP) agents expose standardized tools and data sources through the MCP specification. They simplify integration across vendors and clients by sharing a unified schema and transport layer. Choose an MCP agent when:

You need to deliver capabilities across multiple MCP-compliant clients.
The team values strong governance over context windows, tool calls, and audit logs.
You plan to publish the agent for reuse by partner ecosystems.

Partner Agents

Partner agents are built and maintained by strategic collaborators who co-market or co-support the solution. They often integrate proprietary datasets or industry expertise. Adopt a partner agent when:

You want to leverage a partner's domain authority or certifications.
The go-to-market plan depends on shared support and success motions.
Joint customers expect a pre-integrated experience across both platforms.

Third-Party Agents

Third-party agents come from independent builders or marketplace contributors outside of formal partnerships. They expand the gallery with niche use cases and experimental ideas. Engage a third-party agent when:

You need to fill a gap quickly while evaluating long-term fit.
The risk profile allows for lighter-weight validation or pilots.
You intend to iterate with the creator to customize functionality.

Evaluate agent types together to build resilient solutions: combine orchestrators with utility agents, pair super agents with MCP connectors, and mix partner or third-party options to cover specialized requirements.