AGENT INFRA

AI Agent Infrastructure Projects.

A curated, vendor-neutral directory spanning agent runtimes, sandboxes, memory, identity, observability, durable workflows, and protocols.

Explore and contribute to the AI agent infrastructure project directory on GitHub

Search project names, Areas, or tags.

Filter by Area

Choose one or more to narrow the list.

4 projects found

01

Brazosd

Specialist agents and governed cloud workflows

A platform for running specialized agents in the cloud with event-driven multi-agent workflows, approvals, and enterprise security controls.

Areas

  • 🤝 Communication & Coordination
  • 🔑 Identity & Trusted Tools

Tags

  • Specialist agents
  • Multi-agent
  • Approvals
  • Security
  • Cloud execution
View project

02

Ephemeral Sandbox

Open-source agent sandbox infrastructure for parallel coding agents

An open-source agent sandbox runtime that gives parallel coding agents isolated copy-on-write workspaces over shared durable context, with namespace execution, observability, and controlled atomic publication.

Areas

  • 🧪 Execution & Sandboxing
  • 🧠 Memory, Data & Filesystems
  • 🔎 Observability & Provenance

Tags

  • Agent sandbox
  • Coding agents
  • Multi-agent
  • Copy-on-write
  • Namespace isolation
  • Atomic publication
View project

03

NervaFS

A governed filesystem workspace for agents

A local-first workspace that brings agent identity, memory, tools, data, communication, and audit into one governed environment.

Areas

  • 🧠 Memory, Data & Filesystems
  • 🔑 Identity & Trusted Tools

Tags

  • Filesystem
  • A2P
  • Identity
  • Memory
  • Audit
View project

04

ToWow

An open protocol for agent discovery and collaboration

An open protocol that lets agents discover one another, negotiate roles, and collaborate through semantic matching and structured multi-agent coordination.

Areas

  • 🤝 Communication & Coordination

Tags

  • A2A
  • Agent discovery
  • Semantic matching
  • Coordination
  • Open protocol
View project