speakeasy-api/speakeasy
Manage MCPs, Risk Policies and explore logs in your favorite agent.
Use when migrating MCP servers already configured in a local agent client (Claude Code, Codex, Cursor, Gemini CLI, VS Code or another client) to an explicit Speakeasy AI Control Plane project as remote MCPs, including discovering which servers are missing and optionally restricting them to the organization's Tailscale private network.
Add a reviewed MCP Catalogue server to an explicit AICP project through the Speakeasy AI Control Plane Platform MCP.
Add a user-supplied remote MCP server URL to an explicit AICP project through the Speakeasy AI Control Plane Platform MCP.
Create or update reusable signals and message sensors in an explicit Speakeasy AI Control Plane project, with confirmed atomic authoring and live verification.
Configure the address and network access of an existing MCP server or gateway in an AICP project, then check whether its plugin package reflects the new address.
Use when an administrator of a Speakeasy AI Control Plane project says what they want to catch or prevent, in the dashboard's Guardrails terms or their own words, rather than which detectors to enable. Creates one risk policy in an explicit project from a use-case preset or a bespoke prompt-based guardrail through the Platform MCP.
Put a newly published tool in front of people by adding it to one exact MCP server in an explicit AICP project, creating a new server from the project's function tools when none fits, or take a tool back off, with confirmed and version-protected changes through the Speakeasy AI Control Plane Platform MCP.
Create or select a narrow custom AICP MCP access role, assign it to an explicitly selected member, and verify local desired state while separating unverified provider synchronization.
Inspect and safely change who a risk policy targets in an explicit Speakeasy AI Control Plane project, with confirmed administrator audience changes and clear refusals for unavailable self-exclusion.
Redeploy one Platform-managed MCP server from an explicit source AICP project to an explicit target project in the same organization, complete secure setup there, verify readiness, distribute it, and optionally retire the source with a safe cutover through the Speakeasy AI Control Plane Platform MCP.
Review a Shadow MCP target, approve an explicit audience, and safely onboard and distribute it through the Speakeasy AI Control Plane Platform MCP.
Review proposed improvements to the skills in an explicit Speakeasy AI Control Plane project, then approve, partially approve, correct, or dismiss each one with the user's confirmation. Use when someone asks which skills have suggested changes waiting, wants to work through the suggestion queue, or wants to accept or reject feedback-driven edits to a skill.
Summarize critical Watchdog findings for an explicit Speakeasy AI Control Plane project. Use for a daily security digest or a review of critical findings detected in the last 24 hours, without configuring delivery or scheduling.