AI GOVERNANCE
AI governance for the assets your teams actually use
Most writing on AI governance covers policy frameworks and model risk. Those matter, but they leave a gap if you cannot control the skills, prompts, MCP servers, secrets and agents your people use every day. Joyst is an AI governance platform for those everyday tools. It gives you permissions, folders and an audit trail. It is not a general compliance suite or a developer portal.
WHAT WE GOVERN
An AI governance platform for the tools your teams use.
- Skills, prompts, MCPs, secrets, agents
Joyst governs the skills, prompts, MCP servers, secrets and agents your teams use every day. A policy document or a model-risk checklist does not cover them on its own.
- Controls built in
Joyst puts permissions, folders and an audit trail on every item in the catalogue. It is not a general compliance suite.
- Policies still matter
Responsible AI programmes and compliance frameworks still matter. They leave a gap when nobody controls the skills, prompts, MCP servers, secrets and agents in daily use.
PERMISSIONS
Permissions that match how teams work
- View, use and manage
View, use and manage levels match real work. Groups and folder-wide permissions keep large catalogues manageable, and drafts stay private until they are ready.
- Groups and folders
Grant rights on items and folders to the right groups. Access grows with the team, without a ticket queue.
- One model for everything
The same permissions cover prompts, skills, MCP servers, secrets and agent packages, so control stays consistent from tool to tool.
AUDIT
Audit trail
- Who changed what
See who changed which prompt, skill, MCP server, secret or agent package, instead of hearing that someone updated it in Slack.
- One place to review
Review changes across the private catalogue, so operations and security teams see the same record.
- Tied to permissions
The audit trail sits alongside view, use and manage rights. Who can change an item and who did change it stay linked.
APPROVED USE
Approved catalogues
- Only what is signed off
Everyday reuse stays within the prompts and skills your organisation has approved and published.
- Approved MCP servers
Approve the MCP servers your team and agents may use, instead of unmanaged one-off connections.
- Secrets kept private
Secrets are designed to stay out of the model’s view, so packages remain within approved use.
AGENTS
Agentic AI governance
- Govern packaged agents
Governing agents means controlling what each package includes and who can publish or use it. That covers MCP tools and how secrets are referenced.
- What agents can call
Choose which approved skills, prompts and MCP servers an agent package may use, instead of adding one-off parts for each tool.
- Built for packages
Joyst governs packaged agents and what they include. It does not orchestrate multi-agent workflows.
BESIDE YOUR IDP
Beside your IDP. Not instead.
- Different catalogues
Developer portals catalogue software services and scorecards. Joyst catalogues skills, prompts, MCP servers, secrets and agents.
- Use both
Platform teams keep their IDP for services. Joyst sits beside it for the AI tools those services and agents use.
- Your portal stays
Joyst does not replace Backstage, Port, OpsLevel or Cortex. It works alongside them.
RELATED
Keep exploring.
- Prompt management
A team prompt library with versioning and sharing levels.
Open - Skills registry
Manage, version and move skills across AI tools.
Open - MCP catalogue
Connect, document and approve MCP servers.
Open - Secrets vault
Keys kept out of model context.
Open - Agents
Package skills, prompts, MCP servers and secrets into reusable setups.
Open - Beside your IDP
How Joyst works beside your IDP.
Open - Enable AI safely
Roll out approved AI with owners, permissions and an audit trail.
Open - AI agent security
Controls to put in place before an agent runs.
Open - AI inventory
Record the skills, prompts, MCP servers and agents your teams use.
Open - AI governance framework
Five practical controls, with a checklist.
Open
AI governance FAQ
An AI governance platform controls permissions and keeps an audit trail for the AI tools your teams use: skills, prompts, MCP servers, secrets and agents. Joyst does this in a private catalogue for your organisation.
Developer portals catalogue software services. Joyst catalogues the AI tools your teams use. The two do different jobs and work side by side.
Each item has sharing levels and view, use and manage rights. A central audit trail records who changed what, and folders and groups make the model work at scale.
It means governing packaged agents and what they can include or call. Joyst does not replace your agent orchestration tools.
No. Our core product is not a responsible AI programme or a model-risk framework. We govern the everyday AI tools those programmes often leave uncontrolled.
Access is none, view, use and annotate, or manage. An item can be private, shared with a group, or visible to the organisation. The author and organisation admins can manage it. Organisation-wide visibility gives everyone else view access. Use and annotate is for notes on MCP tools, not for prompts. Admins can use secrets, agents and the MCP catalogue. Members have restricted access. Other permissions are set through user groups.
Organisation admins can open a log of changes. Each entry shows who did it, the action, the resource, a short detail, and the time. It covers prompts, skills, agents, MCP connections, secrets, permissions, folders, groups, tokens and organisation changes. Reads are not logged. It is not a before-and-after copy of the content. Each item also has its own version history.
Govern the AI your teams use.
Create an organisation account and bring skills, prompts, MCP servers, secrets and agents into one private catalogue.
- Private to your organisation
- Draft, then publish
- Limited open beta