AgencyOS
For marketing agencies: 31 AI specialists, a content calendar and a client portal from request to invoice.
Stackmint documentation
Set up and run your AI operating system, and build on the Stackmint platform. Start from your OS, review the governance layer, or go straight to the platform reference.
Find the right place to start.
Each operating system runs on the same platform and the same governance layer. The reference below covers both.
LegalOS
For law firms: 28 AI specialists for contract review, research and drafting, with lawyer approval of the exact version.
Governance and security
Human approval, audit logs, kill switches and client data isolation, under every OS.
Build on the platform
The API, MCP tools and the building blocks behind every Custom OS.
Search docsBrowse the docs
Browse the documentation
Workspaces, specialists, workflows, governance controls and the API. Under the hood, specialists are agents (Dendrites) that run workflows (Branches) made of steps (Buds).
Core Concepts
Workspaces
Workspaces are the top-level isolation boundary in Stackmint. Every user, Capability, budget, and governance policy lives inside a Workspace.
ReadBuds
Buds are the atomic execution units of the Stackmint platform. Each Bud is a typed, versioned function with a defined input schema, output schema, and implementation.
ReadBranches
Branches are execution pipelines that chain Buds together into a governed workflow. They define inputs, outputs, sequencing, and execution constraints.
ReadCapabilities
Capabilities are Workspace-deployed instances of Branches with governance policies, budgets, and access controls applied via a Manifest.
ReadContext & Variables
Context Variables are Workspace-level key/value pairs injected into Branch prompts and Bud inputs at execution time. They decouple configuration from workflow logic.
ReadRenderers
Renderers are certified UI components that format and display Branch execution outputs in Stackmint Surfaces. They are installed per Workspace from the marketplace.
ReadSurfaces
Surfaces are the delivery mechanisms through which Capabilities are exposed to end-users. Stackmint provides Chat, Client Cockpit, Partner Cockpit, and Embedded Surface types.
ReadLonger reads
Guides on running AI work inside a services firm.
Intelligence vs. Execution
Learn how to separate model inference from production execution with gateway controls, deterministic runtime blocks, and governed capability architecture.
Read guideThe SI Survival Playbook
See how systems integrators shift from implementation projects to managed AI-native operations and protected partner IP.
Read guideBecoming AI-Native
Follow the operating playbook for agencies and systems integrators moving from billable hours to recurring governed AI capabilities.
Read guideAgency Monetization Playbook
Learn how agencies and systems integrators package, price, bill, and monetize governed AI capabilities across client workspaces.
Read guideTerms used in these docs
Short definitions, from buyer terms such as human approval and audit logs to platform terms such as Buds, Branches and Dendrites.