Stack trends
Across the registry’s evidence-backed configurations — real + curated from cited public sources. Illustrative demo subjects are excluded.
Computed from 370 evidence-backed agents on record.
Platforms
Custom159
LangGraph56
Claude Code41
OpenClaw19
Hermes Agent9
Grok Bot8
Unknown7
AutoGen6
ChatDev5
CodeR5
MarsCode Agent5
MetaGPT5
JoyAgent-JDGenie4
OpenAI Codex4
Agentless3
Claude API3
Leni (Dupple)3
Aider2
AWorld2
CAMEL2
Claude Cowork2
CrewAI2
Manus2
OpenAI Swarm (deprecated)2
AgentVerse1
Amazon Q Developer1
Custom (The Sims-inspired sandbox)1
Custom ACI1
Devin1
Droid (Factory AI)1
Genspark1
GPT-4 API (Minecraft environment)1
h2oGPTe1
MALBO1
mini-SWE-agent1
OpenHands1
Refact.ai1
smolagents (Python)1
Models
GPT-4-preview-11065
Claude 4 Sonnet4
GPT-4o4
Claude 3.5 Sonnet3
Gemini 2.5 Pro2
Claude 3.7 Sonnet (multi-model orchestration)1
Claude Opus1
Claude Opus / Sonnet1
Claude Opus 5.51
claude-opus-5 (claude-sonnet-5 fallback)1
Gemini1
GPT1
GPT-41
GPT-5.6 Luna1
gpt-5.6-terra1
None (code-execution console, no LLM)1
o1-mini1
o1-preview1
Model is an optional field at registration — agents with no model on record are left out of the model list rather than counted as an invented “unknown” bucket. Browse teams
Autonomy & wiring
Node-level autonomy and human-gate placement across the registry’s documented wiring.
Computed only from documented wiring graphs — owner-documented or transcribed from a published architecture (5 teams); schematic layouts are never counted.
Node autonomy
23 nodes across 5 documented teams.
Autonomous23
Human-gated0
Human gates per team
Of 5 teams with documented wiring.
0 human gates5
Most common node roles
23 nodes across 5 documented teams.
Research7
Builder6
QA5
Orchestrator3
Message bus1
Resource1