Observatory/Agent categories

Agent categories

Explore projects by their purpose, capabilities, and technical direction.

Data as of 2026-11-06

Classification coverage

Uses current active topic assignments for repositories already in the radar on this date.

In the radar
18
Classified
10
Unclassified
8
Coverage
55.6%
About classification

Projects may belong to several types. Automatic suggestions use descriptions and GitHub topics; a manual choice takes precedence. Unclear projects remain unclassified. Coverage refers to this library, not all of GitHub.

View category growth data

15 topics

A parent topic includes projects assigned directly to it and to its child topics. Unassigned projects remain visible from Project monitoring.

TopicParentRepositoriesStars1 day7 days30 days
Coding

AI assistants that write, review, debug, or maintain software.

Top level 2 181,888 N/A0/2 comparable N/A0/2 comparable N/A0/2 comparable
Research

Agents that search, investigate, synthesize sources, and produce research.

Top level 0 N/A N/A0/0 comparable N/A0/0 comparable N/A0/0 comparable
Computer use

Agents that operate websites, browsers, desktop applications, and computers.

Top level 1 14,379 N/A0/1 comparable N/A0/1 comparable N/A0/1 comparable
Workflows

AI agents for business processes, personal tasks, and repeatable automation.

Top level 0 N/A N/A0/0 comparable N/A0/0 comparable N/A0/0 comparable
Agent platforms

Frameworks and workspaces for building, coordinating, and supervising agents.

Top level 0 N/A N/A0/0 comparable N/A0/0 comparable N/A0/0 comparable
Clients & remote access

Desktop, mobile, remote-control, and distributed agent interfaces.

Agent platforms 0 N/A N/A0/0 comparable N/A0/0 comparable N/A0/0 comparable
Multi-agent

Collaboration, delegation, routing, and coordination among AI agents.

Agent platforms 0 N/A N/A0/0 comparable N/A0/0 comparable N/A0/0 comparable
Workbench

Interactive environments for building, supervising, and debugging agents.

Agent platforms 0 N/A N/A0/0 comparable N/A0/0 comparable N/A0/0 comparable
General agents

General-purpose AI agents or agents whose product type is not yet established.

Top level 4 109,341 N/A0/4 comparable N/A0/4 comparable N/A0/4 comparable
Infrastructure

Reusable models, inference engines, data systems, and agent components; these are not necessarily agent products.

Top level 4 504,819 N/A0/4 comparable N/A0/4 comparable N/A0/4 comparable
Tools & connectors

Browser control, computer use, connectors, and tool integrations.

Infrastructure 1 98,610 N/A0/1 comparable N/A0/1 comparable N/A0/1 comparable
Harness

Agent runtimes, execution loops, sandboxes, and orchestration shells.

Infrastructure 0 N/A N/A0/0 comparable N/A0/0 comparable N/A0/0 comparable
Memory

Persistent context, retrieval, knowledge graphs, and agent memory.

Infrastructure 0 N/A N/A0/0 comparable N/A0/0 comparable N/A0/0 comparable
Reliability

Evaluation, observability, guardrails, recovery, and verification.

Infrastructure 0 N/A N/A0/0 comparable N/A0/0 comparable N/A0/0 comparable
Skills

Reusable instructions, capabilities, and agent skill packaging.

Infrastructure 3 406,209 N/A0/3 comparable N/A0/3 comparable N/A0/3 comparable