Hosted Taskade MCP
Connect Claude, Claude Code, Codex, Cursor, VS Code, ChatGPT, Windsurf, or any MCP client to the hosted Taskade MCP server at taskade.com/mcp.
Connect Claude, Codex, Cursor, VS Code, ChatGPT, or any MCP client to your Taskade workspace through the hosted server at https://www.taskade.com/mcp. You install nothing: add the URL, sign in, and your client can work with projects, tasks, agents, and Taskade Genesis app source.
┌──────────────────────────────┐
│ Your MCP client │
│ Claude · Codex · Cursor │
└──────────────┬───────────────┘
│ 1. Add the URL
▼
┌──────────────────────────────┐
│ Taskade sign-in (OAuth) │
│ or a personal access token │
└──────────────┬───────────────┘
│ 2. Call list_spaces
▼
┌──────────────────────────────┐
│ Your paid workspaces │
│ projects · tasks · agents │
│ automations · app/ source │
└──────────────────────────────┘This is the inbound direction: an AI client reaches into Taskade. If you want Taskade to call another MCP server, see Which Taskade MCP do I want?.
Before you start
Check the plan, your role, and your client before you connect.
Plan
Every paid plan: Starter, Pro, Business, Max, and Enterprise. The plan of each workspace sets its access. A Free workspace does not show in list_spaces, and the response adds hiddenForPlan and a hint to show the user.
Client
Any client that supports remote MCP servers over HTTP. The setup for common clients is below.
Credential
Nothing to prepare for OAuth clients. For a client that only takes headers, a personal access token.
Your workspace role sets what each tool can do:
Browse the workspace
Any member
Write app source with write_file
Editor or above
Project and task tools
The same permission checks as the Taskade app
Start an automation with run_flow
Permission to run automations in the workspace
Read run results with get_flow_runs
Permission to edit automations
Read the role instructions of an agent with get_agent or inspect_space
Permission to edit agents. Other members see the agent name and metadata
Server details
Every client uses the same endpoint and the same two sign-in options.
Server URL
https://www.taskade.com/mcp
Transport
Streamable HTTP (remote). No local install.
Sign-in for interactive clients
OAuth 2.0 with PKCE. The server supports dynamic client registration, so you do not register an app.
Sign-in for header-based clients
Authorization: Bearer <personal access token>
OAuth discovery
https://www.taskade.com/.well-known/oauth-protected-resource/mcp
API catalog (RFC 9727)
https://www.taskade.com/.well-known/api-catalog. It links this server, the REST API v1 spec, and the Action API v2 spec
What you can do
A connected client can read your workspace and make changes in it.
Read data
List and read projects, tasks, fields, agents, automations, templates, and media
Edit tasks
Create, update, complete, move, and delete tasks, with dates, notes, assignees, and fields
Projects and agents
Create, copy, complete, and share projects. Create, update, and prompt agents, and attach knowledge
Automations
Run an automation that is On and starts with the Agent Tool trigger ("When asked"), and read its run history
What the hosted server cannot do
Some work stays in the Taskade app or in another API.
Create or edit an automation
Taskade EVE inside Taskade. To start an automation that is On and starts with the Agent Tool trigger, use run_flow (see Run automations)
Build a new app, or publish an app that is not live yet
Taskade EVE inside Taskade. MCP edits the source of an existing app, and an edit to a live app can go live on its own (see Tips and gotchas). publish_kit_version releases a new version of a kit you already published
Set who can use an app
Taskade EVE inside Taskade
Upload or delete media
The REST API (media-upload) or the Action API (upload-media, delete-media)
Rename, export, or soft-delete a project
The Action API v2: renameProject, exportProject, and archiveProject. To archive a project as completed, use complete_project
Register webhooks
For a Taskade EVE job, the client can write the request for the user to paste into Taskade EVE.
Connecting your client
Pick your client below. Each section gives the exact config or command, then the sign-in step.
Claude (desktop and web)
Claude adds remote MCP servers as custom connectors. OAuth runs in your browser.
In Claude, open Settings → Connectors.
Click Add custom connector.
Type
Taskadeas the name. Pastehttps://www.taskade.com/mcpas the URL. Click Add.Click Connect, then sign in to Taskade and allow access.
The Claude labels can change between releases. If you do not see Connectors, search the Claude help for "custom connectors".
Claude Code
Install the Taskade plugin. It adds the server and the Taskade skills, which tell Claude which tool fits each job.
The same plugin installs in other agent hosts:
Codex
codex plugin marketplace add taskade/mcp, then codex plugin add taskade@taskade
GitHub Copilot CLI
copilot plugin install taskade/mcp
Gemini CLI
gemini extensions install https://github.com/taskade/mcp
Any other agent can read the Taskade skill at https://www.taskade.com/SKILL.md. Taskade also publishes an agent skills index at https://www.taskade.com/.well-known/agent-skills/index.json, so this command installs the skill:
The skill also tells the agent how to read any Taskade page as Markdown. The agent sends Accept: text/markdown to the normal page URL.
To add only the server, run one command:
In Claude Code, type /mcp, select the Taskade server, and sign in. To share the server with your team, commit this .mcp.json file at the root of the repository:
Codex
Codex connects to the hosted server over Streamable HTTP and signs in with OAuth. Add the server, then sign in:
The login command opens Taskade in your browser. Sign in and allow access. To install the Taskade skills too, use the Codex plugin command in the Claude Code section.
Cursor
Cursor reads MCP servers from ~/.cursor/mcp.json (all projects) or .cursor/mcp.json (one project).
Save the file. Open Cursor's MCP settings, find taskade, and sign in when Cursor asks.
VS Code
VS Code reads MCP servers from .vscode/mcp.json in your workspace. The key is servers, not mcpServers.
You can also add the server from a terminal:
Start the server from mcp.json, and sign in when VS Code asks. Use the Taskade tools in Copilot Chat agent mode.
ChatGPT
ChatGPT connects to a remote MCP server as a custom connector in developer mode. It uses OAuth only.
In ChatGPT settings, turn on developer mode for apps and connectors.
Create a connector. Type
Taskadeas the name andhttps://www.taskade.com/mcpas the server URL.Select OAuth as the authentication. Sign in to Taskade and allow access.
Developer mode is available only on some ChatGPT plans, and the labels change often. If you cannot find it, search the ChatGPT help for "developer mode" and "connectors".
Windsurf
Windsurf reads MCP servers from ~/.codeium/windsurf/mcp_config.json. A remote server uses serverUrl.
If Windsurf does not open a sign-in page, add the headers block from "Other clients" below.
Other clients
Any client that can send a request header connects with a personal access token. This is the config that Taskade copies from Settings → API → MCP connection → Advanced: use a personal access token:
A client that runs only local stdio servers can reach the hosted server through the mcp-remote bridge. It runs OAuth in your browser:
Scripts, CI jobs, and agent frameworks cannot open a browser. Send Authorization: Bearer $TASKADE_TOKEN with a personal access token. See Use a personal token with hosted MCP.
First-time OAuth authentication
An interactive client signs in once, then keeps the token. The client does the OAuth work, and you only sign in and allow access:
Add the server URL to your client.
Follow the browser redirect to Taskade.
Sign in with the Taskade account that can open the target workspace.
Allow MCP access to your account.
Go back to your client. It stores the token.
Do this once for each client.
Confirm it worked
Ask your client for a first read. A good first prompt:
You are done when the client calls list_spaces and shows your paid workspaces. The list comes in pages, and a page can be empty while more pages remain. If no page lists a workspace, no workspace of this account has MCP access.
A browser GET (Accept: text/html)
A setup page for people
HEAD, curl -I, or MCP with no token
401 and WWW-Authenticate: Bearer
MCP with an OAuth token or a personal access token
The MCP session
The MCP toolset
The tables below list every tool that the hosted server exposes: workspace, project, task, agent, automation, app-editing, and app kit tools. The server also exposes a taskade://getting-started resource that an MCP client can read as a quickstart.
Most tools mirror an Action API v2 operation. The MCP-native tools are inspect_space, read_project, list_automations, write_file, run_flow, get_flow_runs, list_my_kits, and publish_kit_version.
Every tool states the MCP hints readOnlyHint, destructiveHint, and openWorldHint. A client can use the hints to approve read tools without a prompt and to ask before a write. Three tools state openWorldHint: true, because their effects reach outside the workspace: prompt_agent, run_flow, and publish_kit_version.
The server also adds two Taskade keys to the _meta of a tool. A client that does not know them ignores them.
_meta key
Tools
Use it to
taskade/write: true
Every tool that is not read-only
Ask the user before a write
taskade/billed: true
prompt_agent and run_flow
Ask the user before a call that uses AI credits or counts toward the monthly automation runs
Most tools take a spaceId. Call list_spaces first to find it. That list is plan-filtered and paginated, so a page can be empty when more pages remain. Request the next page until hasNextPage is false.
Browse & app editing
list_spaces
List the workspaces with MCP access that you belong to, with your role in each
inspect_space
Read files and directories in a workspace's virtual filesystem - app/, projects/, agents/, automations/, media/. Use viewRange: [start, end] to return a specific line range from a file
write_file
Create, edit, or delete files in your Taskade Genesis app's app/ directory. It can create a file, replace a substring, insert at a line, apply a unified diff, or delete a file
write_file refuses to delete a file while any app source still imports it. Remove or change those imports first. You can bring a deleted file back from the app's Version History.
Read workspace data
list_folders
The folders in a workspace (id, name). The first entry is the workspace itself
list_projects
Projects in a workspace as structured JSON (id, name) with cursor pagination
list_my_projects
The projects you viewed recently, across workspaces. It needs no spaceId
get_project
One project by id (id, name)
read_project
A project's full contents - task tree with assignees, dates, tags
list_tasks
Tasks in a project (id, text, completed, parent) with cursor pagination
list_blocks
The top-level blocks of a project (id, text, completed) with cursor pagination
list_fields
The custom fields on a project. Use a field id with set_task_field_value
list_project_members
The members of a project (handle, display name)
get_share_link
A project's edit and view share URLs, or null when sharing is off. It returns the check URL only for some workspace types
list_templates
The project templates in a workspace (id, name)
list_media
The media files in a workspace (id, kind)
list_agents
The AI agents in a workspace (id, name) with cursor pagination
list_automations
The automations in a workspace (id, name) with cursor pagination
Task writes
These tools mirror the Action API v2 task operations:
create_task
Add a task to a project, optionally under a parent task
update_task
Edit a task's content
complete_task / uncomplete_task
Check a task off, or reopen it
move_task
Reorder a task or move it under a different parent
set_task_note / delete_task_note
Set or clear a task's note
set_task_date / delete_task_date
Set or clear a task's date
assign_task / unassign_task
Assign a project member to a task, or remove them
set_task_field_value / delete_task_field_value
Set or clear a custom field value on a task
delete_task
Delete a task
If an eligible workspace has a Google Calendar connection, a date from set_task_date can sync the same way as a date that you set in Taskade.
Create & orchestrate
create_project
Create a project in a workspace from Markdown
create_agent
Create an AI agent (name + role instructions)
get_agent / update_agent / delete_agent
Read, update, or delete an agent
prompt_agent
Send a one-shot prompt to an agent and get its text reply
list_conversations / get_conversation
Browse an agent's conversation history - get_conversation includes a Markdown transcript
add_knowledge_project / remove_knowledge_project
Attach or detach a project as agent knowledge
add_knowledge_media / remove_knowledge_media
Attach or detach a media file as agent knowledge. The file and the agent must be in the same workspace
If the new agent must read or change project data, pass "enableProjectTool": true to create_agent.
You can change this setting later with update_agent. This setting also enables the required knowledge support.
Run automations
run_flow
Start one run of an automation now. The automation must be On and start When asked (a manual trigger, labeled Agent Tool in the automation editor). Pass input when the trigger asks for inputs. It returns flowRunId and status at once
get_flow_runs
Read the recent runs of one automation, newest first: status, start and end time, and the failed step and a short error for a failed run
run_flow returns before the run ends. Call get_flow_runs to read the result. A run status is running, waiting, completed, failed, or filtered. A waiting run is paused on an approval step until a person decides, or until the step times out. Its endedAt is null, the same as a running run. A filtered run stopped early on purpose and did not fail. If run_flow returns the status unknown, running, or waiting, do not start the run again. Read it with get_flow_runs.
App kits
An app kit is a Taskade Genesis app that you published so other people can add a copy. These tools release a new version of a kit from your MCP client:
list_my_kits
The app kits that you published from the workspaces you own: name, kit page URL, newest version, installs, price, and whether people can add a copy. It returns the publicSpaceAppId of each kit, with cursor pagination
publish_kit_version
Publish a new version of one kit, the same as Publish version in the kit's Version history. Pass publicSpaceAppId and optional notes (up to 500 characters) that buyers see in the kit changelog
Only the workspace owner can call publish_kit_version. list_my_kits lists kits from workspaces that you own, and it hides a kit whose workspace has no MCP access. Archived and removed kits are not listed.
A paid kit keeps its price. You can publish 5 versions of one kit in an hour.
Project lifecycle
create_project_from_template
Create a project from a template. The template must be in the destination workspace
copy_project
Copy a project that you can read into a workspace. It returns the new project id
complete_project / restore_project
Archive a project as completed, or restore it. Both are idempotent
enable_share_link
Turn on a project's share link and return its edit, view, and check URLs
enable_share_link returns an error if the share link is already on, or if your email is not verified.
The virtual filesystem
inspect_space presents each workspace as a directory tree:
You can write files in the app/ directory only. Taskade rejects write_file against projects/, agents/, automations/, or media/. Use the structured tools in the earlier tables to create projects, edit tasks, and manage agents.
Example workflows
Each prompt below names the tools that the client calls.
Refactor a component with Claude
The client calls list_spaces, inspect_space, then write_file two times: one new component file and one App.tsx update.
Stand up a project and brief an agent
The client calls create_project, add_knowledge_project, and prompt_agent.
Audit your app for accessibility
The client reads your files with inspect_space and shows a report before you approve any edit.
Tips and gotchas
These rules prevent most surprises.
Read before you edit. Ask your client to inspect a file first. An edit against an old view can fail or give an unexpected result.
One workspace at a time. Tools take a
spaceId. Before you give a command, point your client at the right workspace.A write goes live. Each successful
write_filecommits the change and starts a new build. On a live app, the default setting publishes that build by itself. If the app owner chose Manual Updates, click Publish in Taskade to make the change live.prompt_agentruns the agent. It uses the agent's model and consumes AI credits. Treat it as a real prompt, not a preview. One account can send 60 prompts in 10 minutes, across all agents.
When to use MCP vs. Taskade EVE
Use Taskade EVE inside Taskade. Use MCP when you work in your own IDE or AI client.
Use Taskade EVE when
Use MCP when
You are already inside Taskade
You work in your IDE
You want to build tasks, automations, and an app in one conversation
You script work across projects, agents, and app code
You build from a prompt from start to finish
You want to use a specific model or tool
You prefer a conversation
You want exact, file-level control
Troubleshooting
Match the message that your client shows to a fix.
MCP access is not available on this workspace's current plan.
The workspace is on the Free plan
Upgrade the workspace to a paid plan, or pick a paid workspace.
Access denied: you are not a member of this space.
Your account is not a member of the workspace
Make sure that the workspace ID is correct, or ask for an invitation.
Access denied: your role "…" does not have write access. Requires collaborator or above.
write_file needs the Editor role or above
Ask the workspace owner for the Editor role. A read-only role can still browse.
list_spaces returns an empty page
The list comes in pages, and it hides workspaces without MCP access
Request the next page until hasNextPage is false.
You do not have permission to run automations in this workspace.
Your role cannot run automations in this workspace
Ask the workspace owner for a role that can run automations.
"…" is Off. Turn it On in Taskade, then run it again.
run_flow starts only an automation that is On
Turn the automation On in Taskade, then call run_flow again.
run_flow starts only an automation that starts "When asked" (a manual trigger).
The automation starts on another trigger
Change the trigger to Agent Tool (the "When asked" trigger), or let the trigger start the automation.
Too many agent prompts in a short time, so the agent did not run. Try again in about … min.
The account sent 60 prompt_agent calls in 10 minutes
Wait the number of minutes in the message, then call prompt_agent again. Do not call it in a loop.
Failed to prompt agent, with the error code credits_exhausted
The workspace has no AI credits left
Upgrade the plan in Taskade, or add credits on a paid plan. A retry does not help.
Too many runs of "…" in a short time, so nothing ran. Try again in about … min.
The account started this automation 30 times in 10 minutes
Wait the number of minutes in the message, then call run_flow once. Do not call it in a loop.
Nothing ran. This workspace used all of its … for this month on the … plan.
run_flow reached the monthly limit of automation runs or steps for the plan
Wait for the limit to reset, or upgrade the plan.
…, so "…" did not run., with the error code connector_plan_required
A step of the automation uses a connector that the plan does not include
Upgrade the workspace plan, or remove that step.
Only the workspace owner can publish a kit version.
publish_kit_version needs the Owner role in the workspace of the kit
Ask the workspace owner to publish, or to run the tool from their account.
No live kit listing with this id. An archived or removed listing takes no new version.
The publicSpaceAppId is wrong, or the kit is archived or removed
Call list_my_kits and use a publicSpaceAppId from its result.
Too many versions of this kit in a short time, so nothing was published. Try again in about … min.
You published 5 versions of this kit in the last hour
Wait the number of minutes in the message, then publish once.
The OAuth sign-in starts again and again
The browser is signed in to another Taskade account, or the stored credential is old
Sign in to the right Taskade account in your browser. Remove the Taskade server from the client, then add it again.
401 from a header-based client
The personal access token is wrong or revoked
Create a new token, and update the Authorization header.
Read a tool error in code
A tool error is a result with isError: true. The text block holds the message from the table above. Many refusals also carry structuredContent.error, so a script or an agent can branch without a text match:
class
auth (membership or role), plan (the plan has no access), input (fix the call), refused (the target is not in a state to run), rate (too many calls), limit (a usage or size limit), or error
code
A short code, for example plan_wall, not_member, forbidden, flow_off, not_manual, connector_plan_required, over_limit, rate_limited, credits_exhausted, or too_large
message
The same text as the text block
next
Optional. Text to show the user, for example the upgrade step
retry
no (the same call fails again), safe (you can call again now), or after_wait (call again after retryAfterSec seconds)
retryAfterSec
Optional. The number of seconds to wait before you call again. An after_wait error carries it
An error with no structuredContent.error is a general failure. Read its text.
Set this up with your coding agent
Paste this block into your coding agent to connect it.
Next steps
Which Taskade MCP do I want?: compare hosted MCP, local Workspace MCP, and the MCP Client connector.
Workspace MCP: run a local server that wraps REST API v1.
Run your live app with AI: run a Taskade Genesis app from a chat with Claude.
Action API guide: the API that most hosted MCP tools mirror.
Authentication: personal tokens and OAuth 2.0 in detail.
This page is also published at taskade.com/docs/mcp/hosted-mcp.
Last updated