Rize MCP Server

The Rize MCP server lets AI assistants like Claude and ChatGPT read and write your Rize data — time entries, clients, projects, and tasks — directly from the chat window. It speaks the Model Context Protocol, so any MCP-compatible client can connect.

info

Rize MCP is listed in the major MCP directories including Claude's MCP directory and the Official MCP Registry.

Quickstart

1. Add the connector. In Claude Desktop: Settings → Connectors → Add custom connector. Paste https://mcp.rize.io/mcp as the URL.
2. Authorize. Claude opens a browser to sign in to Rize. Sign in, review the requested permissions, and approve.
3. Try a prompt. Ask Claude "How much time did I track today?" or "Show me my pending time entry suggestions."

That's it. Claude now has 23 tools to read and write your Rize data.

Connection

Setting	Value
Server URL	https://mcp.rize.io/mcp
Transport	Streamable HTTP
Authentication	OAuth 2.0 (authorization code + PKCE)
Discovery metadata	https://mcp.rize.io/.well-known/oauth-authorization-server

Dynamic client registration is supported — clients that follow RFC 7591 (Claude Desktop, Claude Code, Cursor, Zed, etc.) register themselves automatically the first time you connect.

Authentication

The server uses OAuth 2.0 authorization code flow with PKCE. Here's what happens on first connection:

1. Your MCP client asks the server for a client registration.
2. The client opens your browser to https://app.rize.io/oauth/mcp/signin to authenticate you.
3. You sign in with "Continue with Google" or magic link — the same way you sign in to Rize normally.
4. Rize shows a consent screen listing what the AI assistant will be able to read and write.
5. After you approve, the browser redirects back to your MCP client with an authorization code.
6. Your MCP client exchanges the code for an access token and uses it on every request.

Access tokens expire after 2 hours. Refresh tokens are long-lived, so you'll only need to re-authorize if you explicitly disconnect or revoke access from Rize's account settings.

Scopes

Scope	What the assistant can do
read	List and read your time entries, clients, projects, tasks, and team membership.
write	Create, update, and delete the above within your workspace. Approve or reject AI-generated time entry suggestions.

The server operates entirely within your own Rize workspace. It cannot access other users' data.

Setup per client

Claude Desktop / Claude.ai

1. Open Settings → Connectors.
2. Click Add custom connector.
3. Enter:
   • Name: Rize
   • URL: https://mcp.rize.io/mcp
4. Click Add. The Rize connector appears in your connector list.
5. Click Connect next to Rize. A browser window opens.
6. Sign in with Continue with Google or your Rize magic link, then approve the permissions.
7. Return to Claude — the Rize tools are now available in any chat.

Claude Code

1. Register the server:

   claude mcp add --transport http rize https://mcp.rize.io/mcp

   Or drop a .mcp.json in your project root:

   {
     "mcpServers": {
       "rize": {
         "type": "http",
         "url": "https://mcp.rize.io/mcp"
       }
     }
   }

2. Start (or restart) a Claude Code session so it picks up the new server.

3. Run /mcp — you'll see rize listed with status Needs authentication.

4. Select rize and choose Authenticate. A browser window opens.

5. Sign in with Continue with Google or your Rize magic link, then approve. The page will say Authorization successful — you can close it.

6. Back in Claude Code, run /mcp again to confirm rize is connected. The Rize tools are now ready to use.

Codex (OpenAI)

Add Rize to your project's .mcp.json:

{
  "mcpServers": {
    "rize": {
      "type": "url",
      "url": "https://mcp.rize.io/mcp"
    }
  }
}

Run codex and authenticate when prompted. Rize tools are available immediately after authorization.

ChatGPT

1. Open Settings → Apps.
2. Search for Rize or click Add App.
3. Enter the MCP server URL: https://mcp.rize.io/mcp
4. Authorize with Continue with Google or your Rize magic link, then approve.
5. You'll land back in ChatGPT with the Rize connector attached and ready to use in any chat.

Gemini CLI

Add Rize to your Gemini CLI settings.json (either the global ~/.gemini/settings.json or the project-scoped .gemini/settings.json):

{
  "mcpServers": {
    "rize": {
      "httpUrl": "https://mcp.rize.io/mcp"
    }
  }
}

Start gemini in your project and Gemini will open a browser window for the OAuth flow the first time a Rize tool is called. For the latest Gemini MCP configuration options, see the Gemini CLI docs.

Zed

Zed currently runs MCP servers as stdio child processes and doesn't speak remote HTTP directly. Use the mcp-remote adapter to bridge the two in ~/.config/zed/settings.json:

{
  "context_servers": {
    "rize": {
      "command": "npx",
      "args": ["-y", "mcp-remote", "https://mcp.rize.io/mcp"],
      "env": {}
    }
  }
}

mcp-remote opens a browser for OAuth on first run and caches the token locally.

Any other MCP client

Any client that speaks the Model Context Protocol and supports Streamable HTTP with OAuth 2.0 can connect directly. This includes Cursor, Continue, and anything built on the official MCP SDKs. Clients that only support stdio (like Zed today) can bridge via mcp-remote — see the Zed example above.

Point the client at https://mcp.rize.io/mcp — OAuth (including dynamic client registration) is negotiated automatically. A typical config entry looks like:

{
  "mcpServers": {
    "rize": {
      "url": "https://mcp.rize.io/mcp"
    }
  }
}

If your client doesn't support OAuth discovery yet, check its documentation for how to add a remote MCP server — most use some variant of the snippet above. The exact key names (url, httpUrl, transport, etc.) differ per client.

Tools overview

The server exposes 24 tools. Jump to Tool reference for parameters and return shapes.

Time entries

Tool	Purpose
list_my_time_entries	Your time entries for a date range.
list_team_time_entries	All team members' entries (admin only).
get_time_entry	Full details of one entry by ID.
create_time_entry	Log a new entry.
update_time_entry	Edit times, title, entity assignments, or billing.
delete_time_entry	Remove an entry.

AI suggestions

Tool	Purpose
generate_time_entries	Ask the AI to fill in entries for a date range based on your tracked activity.
regenerate_time_entry	Re-run generation on a pending or failed entry, optionally with custom guidance.
approve_time_entries	Accept pending AI suggestions, optionally assigning a client/project/task.
reject_time_entries	Dismiss pending AI suggestions.
approve_tag_suggestion	Accept an AI-suggested client/project/task tag on an entry.

Analytics

Tool	Purpose
get_my_time_allocation	Your hours grouped by client, project, or task.
get_team_time_allocation	Team-wide allocation (admin only).
list_my_apps_used	Your app and website usage for a date range.

Clients

Tool	Purpose
list_clients	List clients and their hourly rates.
create_client	Add a new client.
update_client	Edit a client's name, rate, color, or status.

Projects

Tool	Purpose
list_projects	List projects.
create_project	Add a new project, optionally under a client.
update_project	Edit a project's name, client, color, or status.

Tasks

Tool	Purpose
list_tasks	List tasks.
create_task	Add a new task, optionally under a project.
update_task	Edit a task's name, project, assignee, color, or status.

Team

Tool	Purpose
invite_team_member	Invite a new member by email (admin only).

Tool reference

list_my_time_entries

List the current user's own time entries for a date range. Returns all statuses by default (active, pending, generating, failed). Sorted by start time with client/project/task details and formatted durations.

Parameter	Type	Required	Description
date	string	yes	Start date in YYYY-MM-DD format.
end_date	string	no	End date in YYYY-MM-DD format. Defaults to date for a single-day query.
statuses	array<enum>	no	One or more of active, pending, generating, failed, rejected. Defaults to all non-rejected.
client_ids	array<string>	no	Filter by client IDs.
project_ids	array<string>	no	Filter by project IDs.
task_ids	array<string>	no	Filter by task IDs.
team_id	string	no	Filter by team. Defaults to the user's default workspace.
limit	number	no	Max entries (1–500, default 100).
cursor	string	no	Pagination cursor from a previous response.

Returns { entries, count, has_more, next_cursor }.

list_team_time_entries

List time entries across all team members (team admin only). Non-admins see only their own entries.

Parameter	Type	Required	Description
date	string	yes	Start date in YYYY-MM-DD format.
end_date	string	no	End date. Defaults to date.
creator_emails	array<string>	no	Filter to specific team members by email.
creator_ids	array<string>	no	Filter by identity ID.
statuses	array<enum>	no	See list_my_time_entries.
client_ids	array<string>	no	Filter by client IDs.
project_ids	array<string>	no	Filter by project IDs.
task_ids	array<string>	no	Filter by task IDs.
team_id	string	no	Filter by team.
limit	number	no	Max entries (1–500, default 100).
cursor	string	no	Pagination cursor.

Returns { entries, count, has_more, next_cursor }.

get_time_entry

Fetch a single time entry by ID with full details including client, project, task, billing info, and AI confidence data.

Parameter	Type	Required	Description
id	string	yes	The time entry ID.

Returns the full time entry object, or { error: "NOT_FOUND" } if the ID doesn't exist.

create_time_entry

Create a new time entry with optional client, project, and task assignment. Times must be in ISO 8601 format — convert user-local times to their timezone (available as _user_timezone in every response) before sending.

Parameter	Type	Required	Description
start_time	string	yes	ISO 8601 start time (e.g. 2024-01-15T09:00:00Z).
end_time	string	yes	ISO 8601 end time.
title	string	no	Title for the entry.
description	string	no	Description of work performed.
client_id	string	no	Client to assign.
project_id	string	no	Project to assign.
task_id	string	no	Task to assign.
team_id	string	no	Team. Defaults to the user's default workspace.
billable	boolean	no	Whether this entry is billable.
idempotency_key	string	no	Unique key to prevent duplicate entries on retry. Recommended for all creates.

Returns the created time entry object.

Example

{
  "start_time": "2024-01-15T14:00:00-05:00",
  "end_time": "2024-01-15T15:30:00-05:00",
  "title": "Client review meeting",
  "client_id": "42",
  "project_id": "187",
  "billable": true,
  "idempotency_key": "meeting-2024-01-15-acme"
}

update_time_entry

Update an existing time entry. Supports changing times, title, description, billing, and entity reassignment. Changing team_id clears entity assignments.

Parameter	Type	Required	Description
id	string	yes	The time entry ID to update.
start_time	string	no	New ISO 8601 start time.
end_time	string	no	New ISO 8601 end time.
title	string	no	New title.
description	string	no	New description.
client_id	string	no	Reassign to this client.
project_id	string	no	Reassign to this project.
task_id	string	no	Reassign to this task.
team_id	string	no	Move to this team (clears entity assignments).
billable	boolean	no	Set billable status.

Returns the updated time entry object.

delete_time_entry

Delete a time entry by ID. Works on entries of any status.

Parameter	Type	Required	Description
id	string	yes	The time entry ID to delete.

Returns the deleted time entry object.

generate_time_entries

Generate AI time entries for a date range. Analyzes your tracked apps, websites, and meetings, then clusters them into multiple entries based on natural activity groups. Skips time slots where previous entries were rejected unless retry_rejected is set.

Parameter	Type	Required	Description
start_time	string	yes	ISO 8601 start time.
end_time	string	yes	ISO 8601 end time.
team_id	string	no	Team. Defaults to the user's default workspace.
retry_rejected	boolean	no	If true, also generate entries in time slots that were previously rejected. Default false.

Returns { entries, count, note }. Generation is asynchronous — poll list_my_time_entries with statuses: ["pending", "generating"] to check status.

Rate limit: 15 calls per minute.

regenerate_time_entry

Regenerate AI content for a pending or failed time entry. Useful when generation failed or you want a better title/description. Optionally provide custom instructions.

Parameter	Type	Required	Description
id	string	yes	The time entry ID to regenerate.
instructions	string	no	Custom instructions to guide the AI (e.g. "focus on the meeting with Client X").

Returns the updated time entry object with a note about async completion.

Rate limit: max 3 regenerations per entry, 15 calls per minute.

approve_time_entries

Approve pending AI-generated time entry suggestions, making them active. Optionally assign client/project/task during approval in one operation.

Parameter	Type	Required	Description
ids	array<string>	yes	Time entry IDs to approve (at least one).
client_id	string	no	Assign this client to all approved entries.
project_id	string	no	Assign this project to all approved entries.
task_id	string	no	Assign this task to all approved entries.

Returns { approved, count }.

reject_time_entries

Reject pending AI-generated time entry suggestions. Rejected entries are kept in the database but hidden from active views.

Parameter	Type	Required	Description
ids	array<string>	yes	Time entry IDs to reject.

Returns { rejected, count }.

approve_tag_suggestion

Approve an AI-suggested client, project, or task tag on a time entry. Tag suggestions appear on pending entries as inline recommendations with confidence scores.

Parameter	Type	Required	Description
id	string	yes	The tag suggestion ID to approve.

Returns { id, tag_type, tag_id, tag_name, status, confidence_score, reasoning }.

list_my_apps_used

List the apps and websites the current user has used during a date range. Returns app names, URLs, window titles, and usage durations. This tool is useful for AI assistants that want to understand what work was done during a time period without relying solely on time entry descriptions.

Parameter	Type	Required	Description
date	string	yes	Start date in YYYY-MM-DD format.
end_date	string	no	End date. Defaults to date.
team_id	string	no	Filter by team.

Returns { apps, count } where each app includes name, total duration, and usage details.

get_my_time_allocation

Get the current user's own time allocation summary grouped by client, project, or task.

Parameter	Type	Required	Description
date	string	yes	Start date in YYYY-MM-DD format.
end_date	string	no	End date. Defaults to date.
group_by	enum	no	One of client, project, task. Default client.
client_ids	array<string>	no	Filter by client IDs.
project_ids	array<string>	no	Filter by project IDs.
task_ids	array<string>	no	Filter by task IDs.
team_id	string	no	Filter by team.
statuses	array<enum>	no	Default ["active"].

Returns total hours, billable hours, and a breakdown by grouping.

get_team_time_allocation

Get time allocation summary across all team members (team admin only). Same parameters as get_my_time_allocation, plus:

Parameter	Type	Required	Description
creator_emails	array<string>	no	Filter to specific team members by email.
creator_ids	array<string>	no	Filter by identity ID.

list_clients

List clients with their hourly rates and team associations.

Parameter	Type	Required	Description
statuses	array<string>	no	Default ["active"].
limit	number	no	Max clients (1–100, default 50).
cursor	string	no	Pagination cursor.

Returns { clients, count, has_more, next_cursor }.

create_client

Create a new client. Clients are top-level entities that projects and time entries can attach to.

Parameter	Type	Required	Description
name	string	yes	Client name.
hourly_rate	number	no	Default hourly rate for billing.
color	string	no	Color hex (e.g. #FF5733).
team_name	string	no	Team to associate with. Defaults to the user's default team.

Returns the created client.

update_client

Update a client's name, hourly rate, color, or status.

Parameter	Type	Required	Description
id	string	yes	The client ID.
name	string	no	New name.
hourly_rate	number	no	New hourly rate.
color	string	no	New color hex.
status	string	no	New status (e.g. active, archived).

Returns the updated client.

list_projects

List projects with their client associations.

Parameter	Type	Required	Description
client_id	string	no	Filter by client ID.
statuses	array<string>	no	Default ["in_progress", "completed"].
limit	number	no	Max projects (1–100, default 50).
cursor	string	no	Pagination cursor.

Returns { projects, count, has_more, next_cursor }.

create_project

Create a new project, optionally under a client.

Parameter	Type	Required	Description
name	string	yes	Project name.
client_id	string	no	Existing client ID to associate with.
client_name	string	no	Client name — creates the client if it doesn't exist.
color	string	no	Color hex.
team_name	string	no	Team. Defaults to the user's default team.

Returns the created project.

update_project

Update a project's name, client, color, or status.

Parameter	Type	Required	Description
id	string	yes	The project ID.
name	string	no	New name.
client_id	string	no	New client ID.
client_name	string	no	New client name (creates if missing).
color	string	no	New color hex.
status	string	no	New status (e.g. in_progress, completed, archived).

Returns the updated project.

list_tasks

List tasks with their project and assignee associations.

Parameter	Type	Required	Description
statuses	array<string>	no	Default ["in_progress", "completed"].
limit	number	no	Max tasks (1–100, default 50).
cursor	string	no	Pagination cursor.

Returns { tasks, count, has_more, next_cursor }.

create_task

Create a new task, optionally under a project. Tasks are the finest-grained unit of work and can be assigned to team members.

Parameter	Type	Required	Description
name	string	yes	Task name.
project_id	string	no	Existing project ID.
project_name	string	no	Project name — creates the project if missing.
assignee_email	string	no	Team member email. Defaults to the authenticated user.
color	string	no	Color hex.
team_name	string	no	Team. Defaults to the user's default team.

Returns the created task.

update_task

Update a task's name, project, assignee, color, or status.

Parameter	Type	Required	Description
id	string	yes	The task ID.
name	string	no	New name.
project_id	string	no	New project ID.
project_name	string	no	New project name (creates if missing).
assignee_email	string	no	New assignee email.
color	string	no	New color hex.
status	string	no	New status.

Returns the updated task.

invite_team_member

Invite a new member to a team by email. Sends an invitation email. Requires team admin permissions. Naturally idempotent — re-inviting an existing member returns the existing record.

Parameter	Type	Required	Description
team_id	string	yes	Team ID.
email	string	yes	Email of the person to invite.
name	string	no	Name of the person being invited.
role	enum	no	One of member, admin. Default member.

Returns the created or existing team member.

Rate limits

• 100 requests per minute per user across all tools.
• generate_time_entries: 15 requests per minute (tighter because each call spawns an AI job).
• regenerate_time_entry: 15 requests per minute, with a max of 3 regenerations per entry.

Exceeding a limit returns a RATE_LIMITED error with a retry_after_seconds hint so your assistant knows exactly when to try again.

Error handling

All tool errors return a JSON object with an error field. Assistants should surface the message and retry where appropriate.

Code	Meaning	How an assistant should react
AUTHENTICATION_ERROR	Token missing, expired, or invalid.	Prompt the user to re-authorize the connector.
AUTHORIZATION_ERROR	User lacks permission (e.g. non-admin calling a team-wide tool).	Explain the permission gap; don't retry.
RATE_LIMITED	Rate limit hit.	Wait retry_after_seconds, then retry.
VALIDATION_ERROR	One or more inputs failed server-side validation.	Show messages to the user and correct them.
NOT_FOUND	The requested record doesn't exist or isn't visible.	Don't retry; ask the user for a valid ID.
SUBSCRIPTION_ERROR	Workspace plan doesn't allow this action.	Surface the billing message; don't retry.
HTTP_5xx / INTERNAL_ERROR	Transient server error.	Retry once; if it repeats, surface the message.

Troubleshooting

"Not authenticated" on every call — Your token expired or was revoked. Open the connector settings and sign in again.

OAuth browser opens but doesn't redirect back — Some browsers block automatic redirects to localhost. Try completing the flow in Chrome or Edge.

generate_time_entries returns no entries — Rize only generates entries from tracked activity. If the desktop app wasn't running during the target range, or the range is too short (under 30 minutes), generation may produce nothing. Try a full workday.

"Admin only" errors on team tools — list_team_time_entries, get_team_time_allocation, and invite_team_member require team admin. Use the _my_ variants instead, or ask an admin.

Stale tool list after update — Claude Desktop occasionally caches the tool list. Fully quit and relaunch the app, then re-open a chat.

Example prompts

• "How much time did I track today?"
• "Show me my pending suggestions and approve them."
• "What's my time allocation this week grouped by project?"
• "Create a time entry from 9am to 10am titled 'Team standup'."

Support

• Website: rize.io
• Contact: rize.io/contact
• Privacy policy: rize.io/privacy-policy
• Terms of service: rize.io/terms-of-service