Moira logo

Moira

Runs multi-step, schema-validated AI workflows over MCP with persistent execution state and a web interface.

Install Moira

Pick your client, then add it in one click or paste the config.

Streamable HTTP (Remote)
Server URL
https://moira-mcp.com/mcp

Remote servers are added in the app, not the config file: Customize → Connectors → Add custom connector, then paste this URL.

Hosted endpointLive · TodayLive Remote EndpointOur automated health check connected to this remote MCP endpoint successfully.Last checked 10/27/2026, 12:17:47 AM

Clients with native remote MCP support can connect directly to this URL.

Remote endpoint
https://moira-mcp.com/mcp

Overview

Moira MCP server lets an AI agent discover and run multi-step workflows through an authenticated MCP endpoint. Each step supplies instructions and completion criteria, then checks the agent's response against a schema before persisting the run for later continuation. A web UI supports workflow editing, execution inspection, and account administration. Reach for Moira when agent tasks need repeatable sequencing, guarded continuation, and inspectable execution state.

Use cases

  • Run repeatable multi-step agent workflows
  • Resume interrupted executions in another agent session
  • Inspect workflow runs through a web interface
  • Validate structured step responses against schemas
  • Publish workflow-generated HTML artifacts

Key features

  • Authenticated MCP endpoint
  • Persistent workflow execution state
  • Per-step instructions and completion criteria
  • Schema-checked responses
  • Workflow editing and run inspection UI
  • Guarded continuation and cancellation

Security & health

What it can reach and how reliably it runs — from our own checks, not the maintainer’s claims.

ActiveLast checked Today

Recent health check succeeded.

9/9 checks healthy over the last 43d

Adoption & maintenance

Factual signals from GitHub, npm, and our automated checks — not a rating.

GitHub stars128Stargazers on the source repository.
Last commitTodayMost recent push to the default branch.

How Moira works

What Moira does

Moira coordinates multi-step work for AI agents through the Model Context Protocol. An agent connects to the MCP endpoint, selects a workflow from the catalog, receives the active step's instructions and completion criteria, and submits a structured result. The engine validates that result against the step's schema and records execution state so a later agent session can continue the same run.

The open-source application includes the workflow engine, MCP server, web API, web UI, documentation, and self-host deployment files. A managed service is also available at moira-mcp.com. The UI covers workflow editing, run inspection, and account administration.

How it works

Agents first discover available workflows and choose the exact owner/slug for the task. The runtime then guides the preparation and execution process, including step attempts, continuation, and guarded cancellation. Runtime help can expose English guidance, while help({ topic: "tools" }) renders the typed tool catalog.

Moira stores application accounts, workflow definitions, and execution records in its application data. Workspace files, complete command output, operation receipts, and immutable published artifacts remain in the associated workspace service or paired controller instead of the application database.

Setup and configuration

Self-hosting requires Docker with the Compose plugin. Copy .env.example to .env, keep the supplied Traefik configuration beside the Compose file, and start the stack with Docker Compose:

bash
cp .env.example .env
docker compose up -d

The default local deployment exposes the web UI at http://localhost:8080, documentation under /docs/, and the MCP endpoint at /mcp. For another host, configure MOIRA_HOST, MOIRA_PORT, and STATIC_ARTIFACTS_DOMAIN together. Public artifact hosting also needs matching DNS and wildcard TLS configuration.

The first startup generates missing secrets and prints the administrator password once. Self-hosted registration requires administrator approval, and new accounts can be approved from Admin Panel → Users. Preserve data/.secrets.env with the database and back up workspace-owned stores separately.

Tools and capabilities

  • Discover workflows and select a specific owner/slug.
  • Prepare and execute workflow steps.
  • Submit schema-checked step responses.
  • Continue existing runs across agent sessions.
  • Cancel runs through guarded cancellation.
  • Inspect runtime help and the typed tool catalog.
  • Create or edit workflow definitions through the workflow authoring tools and CLI.
  • Publish HTML artifacts with ownership, quota, expiry, and browser-isolation rules.

Limitations and notes

Schema validation checks response structure; it does not establish that the underlying work actually occurred. Database backups do not include workspace file bytes, full command output, receipts, or immutable publications. Those owner-controlled stores require separate backups and recovery procedures.

Optional workspace and extension features have additional constraints. Local VM execution requires a supported macOS Apple-silicon runtime and explicit owner grants. Extensions are disabled by default and run through a separate optional Compose profile. Self-hosted deployments should preserve the supplied ingress upgrade route and configure proxy trust for their topology.

Reviews

No reviews yet — be the first to share how it worked for you.

Frequently Asked Questions about Moira

Moira is an MCP server that connects AI agents to a workflow engine through an authenticated MCP endpoint. Its main tools help agents discover workflows, prepare and execute steps, submit schema-checked responses, continue existing runs, inspect runtime help, and perform guarded cancellation while execution state is persisted.

Install Moira with Docker Compose by copying `.env.example` to `.env` and running `docker compose up -d`. The local deployment exposes the MCP endpoint at `http://localhost:8080/mcp`, along with the web UI and documentation.

Connect the client to your Moira instance's `/mcp` endpoint and complete its authenticated authorization flow. After connecting, ask the agent to discover workflows and select the exact current `owner/slug` for the task.

No. Moira validates the structure of an agent's response against a schema, but that check does not prove the underlying work happened.

Related MCP servers

Add a badge to your READMEShow it’s listed — live SVG, Markdown or HTML.

AllMCPs Directory Badge

Full Badge Customizer

Showcase your server listing on GitHub or your project documentation. Embed this dynamic SVG badge to highlight official listing status and live engagement.

Badge Style:
Live Dynamic SVG PreviewMoira AllMCPs Directory Badge
Markdown (GitHub README)
[![AllMCPs](https://allmcps.com/api/badge/moira?style=directory)](https://allmcps.com/mcp/moira)
HTML Embed
<a href="https://allmcps.com/mcp/moira"><img src="https://allmcps.com/api/badge/moira?style=directory" alt="Moira on AllMCPs" /></a>