Data extraction teams

Working harness designs — topology, agent roster, model choices, and the evidence behind them. 5 teams match.

See stack trends →
SortFeaturedRecently activeHighest trust
🔀

Anthropic Orchestrator-Workers Pattern

Anthropic

Self-ReportedCurated

Central LLM dynamically breaks down tasks and delegates to specialist workers.

updated 1y ago
Orchestrator–Worker2 agentsClaude API
OrchestratorWorker
software-deliveryresearchdata-extraction
Outcome[unknown]
Economics[unknown] · deliberate
1 proof
💬

AutoGen Group Chat

Microsoft Research (AutoGen)

Self-ReportedCurated

Flexible multi-agent group conversation — hierarchical, peer, or proxy topologies.

updated 3y ago
Supervisor3 agentsAutoGen
ConversableAgent
software-deliveryresearchdata-extraction
Outcome69.5%
Economics[unknown] · deliberate
1 proof
🕹️

Magentic-One

Microsoft Research

Self-ReportedCurated

Microsoft Research generalist 5-agent system: GAIA 32.33%, WebArena 32.8%.

updated 1y ago
Supervisor5 agentsAutoGen
Orchestrator· GPT-4oWebSurfer· GPT-4oFileSurfer· GPT-4oCoder· GPT-4o+1 more
researchsoftware-deliverydata-extraction
Outcome[unknown]
Economics[unknown] · deliberate
1 proof
🐍

smolagents CodeAgent

Hugging Face

Self-ReportedCurated

Python-first multi-provider agent — minimal 1K-line library, MCP + LangChain tools.

updated 1y ago
Solo + Tools1 agentsmolagents (Python)
CodeAgent
software-deliveryresearchdata-extraction
Outcome[unknown]
Economics[unknown] · deliberate
1 proof
🌅

Helios Swarm

Helios Labs

Self-ReportedIllustrative

Homogeneous extraction worker pool deployed from a single blueprint.

updated 2mo ago
Swarm3 agentsCustom
Blueprint origin· Claude Haiku 4.5Worker· Claude Haiku 4.5Worker· Claude Haiku 4.5
data-extractionresearch
Outcome[unknown]
Economics2.1M
0 proof