Skip to main content
Product overview

One source of truth for every AI agent your team uses.

Author, govern, distribute, and audit every agent from one place. Versioned like real software. Available in every runtime your team works in.

app.agentbundle.dev
The AgentBundle Org Defaults page showing company context, org-wide skills, MCPs, and guardrails inherited by every agent.
For every team

Agents are shared infrastructure for your whole organization.

AgentBundle treats AI agents like the rest of your software stack — versioned, distributable, and governable across every role. A marketing lead authors an on-brand brief-to-draft agent that engineers use for technical RFCs. A platform engineer authors a release-note agent that marketing uses for launch copy. One canonical definition, the same security scans on every publish, available in every team’s preferred runtime.

Author from any seat

Click New Agent in the dashboard, pick a starting template (or start from scratch), fill in the form. Engineers can scaffold via the apm CLI. Same output either way: one bundle, every runtime.

Install once, available everywhere

Run apm install from the CLI, or download the bundle as a zip from the marketplace. Each agent ships to Claude Code, Cursor, GitHub Copilot, OpenCode, Gemini, Codex, and Windsurf — wherever a user already works.

Governance and audit, by default

Owner, Admin, and Member roles on every plan. Department-level admins, N-required reviewer approval workflow, and full audit-log retention land on Business — auditable end-to-end.

How it works

Three steps from definition to running everywhere.

1

Define

Author your agent once: persona, skills, MCPs, guardrails. One canonical source of truth your whole team can review and edit — from the dashboard or your editor.

The New Agent form in AgentBundle: a form to define an agent's persona, skills, and guardrails.
2

Distribute

Publish, and the platform produces APM-compatible bundles for all 7 runtimes — Claude Code, Cursor, Copilot, OpenCode, Gemini, Codex, Windsurf. Same agent, every target, re-built on every change.

The dashboard listing published agents as cards with their distribution status across runtimes.
3

Use

Anyone in the org installs the agent in their preferred runtime. Engineers run it in Claude Code or Cursor; non-technical teammates click through to a target-shaped zip and drop it in. Same agent everywhere.

An agent detail view in AgentBundle, showing its current version, target runtimes, and one-click install actions.
Built on APM

An open package format for AI agents.

AgentBundle outputs APM-compatible packages — the open packaging spec for AI agents, maintained by Microsoft. AgentBundle is the visual platform that authors and distributes these packages; APM is the standard format and the CLI for installing them. Same artifact, two install paths.

For developers

Run apm install owner/repo/agent in the terminal. Microsoft’s apm CLI auto-detects your IDE, drops files into the right target directories, wires MCPs, and locks the version in apm.lock.yaml.

For everyone else

Open the agent in the AgentBundle dashboard, pick your target, and click Download. Get a zip shaped for your runtime that drops into your project. Identical to what the CLI produces — no terminal needed.

Trust & control

Built for teams that need a paper trail.

Versioning, reviews, audit, and guardrails are first-class — not a tier. The whole org sees every change, every publish, every install.

An agent's versions page showing the version history, status (live, deprecated, recalled), and per-version deprecate, recall, and revert actions.
Versioning

Every change is a version.

Every publish is a semver release. Browse the full version history with diffs, roll back to any prior version, deprecate a version to warn consumers on install, or recall it outright to block installs entirely. Lifecycle management for AI agents the way you'd manage a real software dependency.

An agent's review tab showing unpublished changes, a structured diff of edits, and the Submit-for-Review and Publish controls.
Approvals

Reviews before things ship.

Owner, admin, and member roles on every plan. On Business, add department-level admins and configure N-required reviewers — publishes hold until your reviewers sign off. The reviewer sees a structured diff of every change before approving.

A chronological activity log showing recent agent publishes and policy changes with the actor and timestamp for each.
Audit trail

Who shipped what, when, and why.

Every publish, install, edit, and rollback is recorded with the actor, timestamp, and before/after diff. Filter by member, department, or agent. Export to CSV for your auditor in one click.

The organization guardrails inventory showing org-wide safety and refusal rules applied to every agent on publish.
Guardrails

Org-level safety, applied automatically.

Define refusal patterns, tool restrictions, and content rules once at the org level. Every agent inherits them on publish — no per-agent configuration drift, no missed copy-paste between teams.

The Org Defaults page showing company context, org-wide skills, guardrails, and MCPs that every agent inherits unless explicitly overridden.
Globals & overrides

Set rules once. Override only where it matters.

Define company context, skills, MCPs, and guardrails at the org level — every agent inherits them by default. Each agent can override what it needs (a sales agent gets the CRM MCP, an engineering agent gets GitHub) without touching org defaults. New agents inherit the latest org globals; existing agents pick up changes on next publish.

Stop redoing your best work on everyone's laptop.

Define your team's agent once. Hand it to every teammate and every new hire.

Free for teams up to 5. No credit card.

Copied to clipboard