Skip to content
AgentField — The AI Backend. Build agents like APIs.

Build agents like services. Scale them beyond chatbots — the AI Backend for production agents. Write Python, Go, or TypeScript and get callable REST APIs, async execution, memory, discovery, observability, identity, and verifiable audit, built in.

⭐ Star AgentField · Docs · Website · Discord


Built on AgentField

Production-grade autonomous agents. Apache 2.0. Each independently deployable.

Project What it does
SWE-AF Autonomous engineering backend — hundreds of agents plan, code, test, and ship from one API call
sec-af Autonomous security-audit backend — agents prove exploitability with traces and evidence, no analyst in the loop
plandb Agent-native issue tracker — backend that Claude Code, Codex, and other agents call directly (no human UI required)
af-deep-research Autonomous research backend — agents fan out, recurse, and self-correct without human direction
reels-af Background video pipeline — multi-agent backend ships viral reels / TikToks at ~$0.10 each, lights-out
pr-af Autonomous PR-review backend — agents triage every pull request before a human ever opens it
af-reactive-atlas-mongodb Background MongoDB enrichment — agents watch any collection and add an AI intelligence layer as documents arrive
cloudsecurity-af Autonomous IaC scanner — agents reason about attack paths in the background, no analyst review required
roboscribe-af Autonomous dataset annotator — agent swarm labels LeRobot demos without human raters

All run on agentfield — one binary handling identity, queues, discovery, audit, and routing.

See all examples on agentfield.ai →


Why this org exists

Software is moving from deterministic backends to autonomous backends — fleets of agents reading ledgers, moving money, rotating credentials, coordinating across vendors. That world needs three things AgentField provides:

  • Agents as first-class actors — cryptographic identity per agent, not API keys
  • Verifiable provenance — every decision signs an audit trail you can verify offline
  • Autonomy as a primitive — queues, webhooks, discovery, policy built into the runtime

How AgentField compares to LangChain, CrewAI, Temporal →

Pinned Loading

  1. agentfield agentfield Public

    Build, run and scale AI agents like API and microservices - observable,auditable and identity-aware from day one.

    Go 2.1k 333

  2. SWE-AF SWE-AF Public

    Autonomous software engineering fleet of AI agents for production-grade PRs on AgentField: plan, code, test, and ship.

    Python 834 133

  3. af-deep-research af-deep-research Public

    Autonomous AI backend for deep research AI applications.

    Python 80 19

  4. sec-af sec-af Public

    AI-native security auditor on AgentField that proves exploitability with verdicts, traces, and actionable evidence.

    Python 95 17

  5. af-reactive-atlas-mongodb af-reactive-atlas-mongodb Public

    Turn any MongoDB Atlas collection into an AI-powered intelligence layer. Config-driven, domain-agnostic document enrichment powered by AgentField.

    Python 8 2

  6. plandb plandb Public

    The issue tracker your AI agents are missing. Think Linear or Jira, but for your Claude Code..

    Rust 90 6

Repositories

Showing 10 of 12 repositories

Top languages

Loading…

Most used topics

Loading…