Agent-Field/agentfield

Build, run and scale AI agents like API and microservices

What it solves

AgentField solves the problem of moving AI agents from local prototypes to production-grade infrastructure. It provides a control plane that handles the complex backend requirements—such as routing, queuing, retries, and observability—that typically require manual engineering when scaling agents from a single script to thousands of concurrent, distributed microservices.

How it works

You write your agent logic as standard Python, Go, or TypeScript functions. AgentField wraps these functions, automatically exposing them as REST endpoints and registering them with a central control plane. When your agents need to interact, they use the framework to call one another; the control plane manages the communication, handles asynchronous execution, and tracks the entire workflow in real-time. It supports advanced features like human-in-the-loop approvals, canary deployments, and cryptographic identity for secure, audited agent-to-agent communication.

Who it’s for

It is designed for developers who have built agent prototypes using frameworks like LangChain or CrewAI and are now ready to deploy them as reliable, scalable services. It is ideal for teams needing to manage complex, multi-agent workflows that require production-level governance, observability, and the ability to handle high-volume fan-out requests.

Highlights

  • Production-ready APIs: Automatically turns agent functions into REST endpoints.
  • Scalable Orchestration: Handles thousands of concurrent agents with built-in queuing, retries, and backpressure.
  • Governance & Security: Provides cryptographic identities for agents and generates tamper-proof audit trails for every execution.
  • Observability: Offers real-time workflow DAGs, metrics, and execution traces out of the box.
  • Multi-Language Support: Write agent logic in Python, Go, or TypeScript with no need for complex DSLs or YAML configuration.

Related

  • Project
    eunomia-bpf/agentsight

    AgentSight is a kernel‑level observability tool for AI agents. Using eBPF and TLS tracing, it records prompts, model calls, subprocesses, file and network activity without requiring any SDK or proxy. It provides live dashboards (`top`), replay visualisations (`vis`), queryable SQLite logs, and OpenTelemetry export, helping developers debug failures, audit security, and understand token/resource costs of agents such as Claude Code, Gemini CLI, and Python‑based tools.

  • Project
    AgentEra/Agently

    Agently is a Python framework that turns LLM calls into structured, observable, and restart‑safe runtime units. It provides contract‑first output schemas, streaming of partial results, tool/action integration, versioned skills, managed execution resources, and an event‑driven workflow engine (TriggerFlow) for building reliable AI services.

  • Project
    AgentSystemLabs/agent-office

    Agent Office is a TypeScript‑based tool that visualises LLM‑powered coding agents as desks in a 3‑D office. Each GitHub repo becomes a floor; agents can be spoken to, run terminals, manage git worktrees, and open PRs. The UI runs locally or on a cloud VM accessed via SSH tunnel, with deployment scripts for AWS, Azure, Railway, Fly.io, Dokploy and more.

  • Project
    bytedance/deer-flow

    DeerFlow is an open‑source framework that lets you build autonomous AI agents capable of orchestrating sub‑agents, using tools (web search, sandboxed code execution, RAG), and managing memory. It supports many LLM providers, offers a web UI, a terminal UI, and integrates tracing services, all deployable via Docker.

  • Project
    nullpointexception-i/agent-sphere

    AgentSphere is a Java‑Spring/React platform that runs LLM‑driven ReAct loops, supports multi‑agent delegation, real‑browser automation via a Chrome extension, and a public skill marketplace. It integrates many model providers, offers OIDC SSO, and provides an embeddable chat widget.