Agent-Field/agentfield

Health v2 · maintenance only

Adoption not part of the score

2527 stars · 409 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-30, confidence not recorded

AgentField is an open-source AI backend that lets you build, deploy, and govern AI agents like APIs and scale them like microservices. It provides a Go control plane handling routing, execution queues, retries, memory, agent identity (DIDs/verifiable credentials), access policies, and audit trails, with Python, TypeScript, and Go SDKs.

Use cases

• run thousands of ai agents at scale like microservices
• call ai agents from backend services via rest api
• orchestrate multi-agent workflows with retries and tracing
• add identity and access control between ai agents
• trigger agents from github, stripe, slack or webhooks
• audit every ai decision with cryptographic receipts
• build human-in-the-loop approval workflows for agents
• discover and call agents agent-to-agent mcp style

When to choose

• you need production infrastructure for agents, not just an authoring framework
• you want built-in agent identity, policy, and audit trails
• you need to fan out requests to many agents with queuing and retries
• you want agents triggered by events from your existing stack
• you deploy on kubernetes or cloud-native environments

When to avoid

• you only need a lightweight in-process agent library for a prototype
• you want a simple prompt-chaining framework without infrastructure
• your stack cannot run a separate control plane service
• you need only a single agent with no scaling or governance needs

Facets

framework · maturity active

agent-framework api-framework http-server rpc auth authorization workflow-automation webhook mcp rag monitoring tracing scheduling service-discovery sdk large-language-models artificial-intelligence backend microservices developer-tools self-hosted security go python self-hosted cloud cross-platform cli ai-backend agent-orchestration agent-identity decentralized-identity verifiable-credentials audit-trail human-in-the-loop multi-agent-systems control-plane cloud-native agent-scaling observability ai-agents nodejs docker kubernetes

Member repositories

For agents

markdown · JSON · MCP: product_card(name="Agent-Field/agentfield")