Connecting AI Assistants to Manifestly (MCP Server)

What is the Manifestly MCP Server?

The Manifestly MCP (Model Context Protocol) server lets you connect AI assistants — like Claude, ChatGPT, and other MCP-compatible clients — directly to your Manifestly account. Once connected, you can manage your workflows and runs using natural language conversations.

With the MCP server, your AI assistant can:

  • Check your workload and show overdue assignments
  • Start workflow runs by name
  • Complete, skip, or assign steps
  • Fill in data collection fields
  • Browse workflows, runs, departments, and users
  • Add comments to steps
  • Manage webhooks

Supported AI Clients

The MCP server works with any AI client that supports the Model Context Protocol, including:

  • Claude (claude.ai) — Anthropic's AI assistant, via the Connectors feature
  • Claude Code — Anthropic's CLI tool for developers
  • ChatGPT — OpenAI's assistant, via MCP connector support
  • Other MCP-compatible tools and IDE extensions

Setting Up the Connection

Claude (claude.ai)

  1. Open Claude at claude.ai
  2. Go to Settings and look for the Integrations or Connectors section
  3. Search for Manifestly and click to add it
  4. You'll be redirected to Manifestly to log in and authorize the connection
  5. Once you see the "Connected!" confirmation, return to Claude and start chatting

Claude Code (CLI)

Add the Manifestly MCP server to your Claude Code configuration:

  1. Open Claude Code settings (or your .claude/settings.json file)
  2. Add Manifestly as an MCP server with the URL: https://mcp.manifest.ly/mcp
  3. When you first use a Manifestly tool, you'll be prompted to authorize via your browser

ChatGPT

  1. In ChatGPT, go to Settings > Connectors
  2. Add a new MCP connector with the URL: https://mcp.manifest.ly/mcp
  3. Complete the OAuth authorization flow when prompted

What You Can Do

Once connected, just ask your AI assistant in natural language. Here are some common things you can do:

Check your workload

Ask: "What needs my attention?" or "What's on my plate today?"

The assistant will show your overdue assignments, late runs, and in-progress work in one summary.

Start a workflow run

Ask: "Start the daily store opening checklist" or "Kick off onboarding for Jamie Rivera"

The assistant finds the workflow template by name and creates a new run. You can also specify a due date, participants, or a custom title.

Track progress on a run

Ask: "How's the security audit going?" or "Show me the status of Jamie's onboarding"

The assistant retrieves the run, shows step completion status, and surfaces any comments left by team members.

Complete or skip steps

Ask: "Mark the first step as done" or "Skip the visual inspection step"

The assistant completes or skips the step and shows you what's next.

Assign steps to team members

Ask: "Assign Jordan Lee to the next incomplete step"

Just provide the person's name or email — the assistant resolves the user and assigns the step automatically.

Fill in data collection fields

Ask: "Fill in Dana Park's onboarding details — she's an Engineer starting next Monday"

The assistant discovers the available fields, fills in the values you provide, and completes the steps automatically.

Browse your workflows

Ask: "What workflows do we have?" or "Show me the steps in our onboarding workflow"

The assistant lists your workflow templates and can display the full step structure including sections and substeps.

Add comments

Ask: "Add a comment to the access review step: Found 2 stale accounts"

The assistant adds your comment to the specified step.

Manage webhooks

Ask: "List our webhooks" or "Create a webhook for run completed events"

The assistant can list, create, update, and delete webhooks for your account.

Authorization and Security

The MCP server uses OAuth 2.0 to securely connect to your Manifestly account:

  • Your credentials are never shared with the AI client. You log in directly to Manifestly and grant permission.
  • Tokens are encrypted at rest using AES-256-GCM encryption.
  • Each AI client gets its own isolated session. Connecting Claude doesn't affect your ChatGPT connection, and vice versa.
  • You can revoke access at any time from your Manifestly account settings under Connected Apps.
  • The MCP server acts only on behalf of your authenticated user — it respects your existing roles and permissions.

Managing Connected Apps

You can view and manage your MCP connections from your Manifestly account:

  1. Go to your Account Settings
  2. Look for Connected Apps or Authorized Applications
  3. You'll see each AI client that has been authorized (e.g., "Claude", "ChatGPT")
  4. Click Revoke to disconnect a specific AI client

Troubleshooting

"Unauthorized" errors

If your AI assistant reports an authorization error, your session may have expired. Ask the assistant to reconnect, or re-authorize from your AI client's settings. The assistant will provide a link to complete the authorization.

"Not found" errors

The MCP server respects your Manifestly permissions. If you can't access a workflow or run through the AI assistant, verify that you have access to it in Manifestly. Department admins can only see resources in their departments.

The AI assistant doesn't know about Manifestly

If the assistant doesn't seem to recognize Manifestly commands, make sure:

  • The Manifestly connector/integration is enabled in your AI client's settings
  • You've completed the authorization flow
  • Try asking "What can you do with Manifestly?" — this triggers the about tool, which explains available capabilities

Data not showing up

The MCP server reads data in real-time from your Manifestly account. If something seems missing:

  • Check that the data exists in Manifestly's web interface
  • Verify you have the right permissions (admin features require admin access)
  • For user lookups, admin access is required

Available Tools Reference

The MCP server provides 39 tools organized into categories:

Reading Data

  • Verify Authentication — Check your connection status
  • List/Get Workflows — Browse workflow templates and their steps
  • List/Get Runs — View workflow runs filtered by status, workflow, department, or tags
  • Get Run Step — View details of a specific step
  • List Run Step Comments — View comments on a step
  • Get Fields — View data collection field definitions and current values
  • List My Assignments — View steps assigned to you (or another user, for admins)
  • List/Get Departments — View departments and their members
  • List/Get Users — View user details (admin only)
  • List Webhooks — View registered webhooks

Taking Action

  • Create Run — Start a new workflow run
  • Update/Close Run — Modify or close a run
  • Complete/Uncomplete Step — Mark a step as done or undo it
  • Skip/Unskip Step — Skip a step or un-skip it
  • Start/Stop Step — Track time on a step
  • Assign Step — Assign a step to a team member (by ID or email)
  • Update Step Due Date — Change when a step is due
  • Add Comment — Add a comment to a step
  • Create Step — Add a new step to an existing run
  • Set Fields — Fill in data collection field values
  • Import Workflow — Import a workflow from a structured definition
  • Create/Update/Delete Webhook — Manage webhook subscriptions

Smart Tools

  • Get My Status — One-call summary of your overdue assignments, late runs, and in-progress work
  • Start Run by Workflow Name — Find a workflow by name and start a run in one step
  • Assign Step by Email — Assign a step using just an email address
  • About — Learn what the connector can do and get usage tips

Need Help?

If you're having trouble with the MCP server connection, contact our support team through the help widget in the bottom-right corner of any Manifestly page, or email support@manifest.ly.

Still need help? Contact Us Contact Us