MCP App Store
by KeenEthics
Project Simple icon

Project Simple MCP Server for ChatGPT

By Project Simple

Overview

Project Simple MCP connects AI assistants and agents directly to your product and engineering workspace. Manage project roadmaps, goals and milestones, initiatives and epics, backlogs, sprints, and teams through a unified interface. Retrieve real-time delivery information, analyze progress, identify risks, and understand what teams are working on and when work is expected to be completed. AI agents can interact with work items and tasks and use Project Simple's built-in analytics to answer questions about delivery predictability, quality, team performance, and project health. With Project Simple MCP, your AI works with the same product and delivery context as your teams, connecting strategy, planning, and execution without requiring manual status reporting or disconnected tools.

How to connect

  1. ChatGPT: Open the Project Simple listing in the ChatGPT app directory and choose Connect. ChatGPT listing Project Simple exposes 137 tools here.

Tools

add_epic

ChatGPT
Create a new epic in a project. Use it when the user asks for a body of work to be grouped under an epic; check with them first if the name has to be invented rather than quoted. Name is required; description, size, label and team are optional and can be set later. Creating an epic does not create an item. Items are filed under the epic afterwards with move_item_to_epic, not here. Returns the new epic in full.

add_epic_comment

ChatGPT
Post a comment on an epic as the signed-in user. Use it to record a decision, answer a question about the epic or leave a note the team will see - it is visible to everyone who can see the epic, so show the user the text before posting anything they did not dictate. Pass replyToCommentId to reply to an existing comment on the same epic. For @Name or @@Name, first search with find_epic_comment_mentions. If a person is found, confirm the choice and pass it in mentions. If no person is found, post the @ text normally without a mention notification. Returns the comment as it was posted.

add_epic_comment_file

ChatGPT
Attach one file to a comment on an epic - not to the epic itself; use add_epic_file for that. The file still shows up in list_epic_files, which lists the files on the epic's comments along with the epic's own. The file's bytes are passed base64-encoded, and they pass through the conversation, so this only suits small files, up to 2 MB; a larger file has to be uploaded in Project Simple, or attached to the item, epic or TODO item itself with start_file_upload. Returns the attached file's id, name, media type and size.

add_epic_comment_with_mentions

ChatGPT
Deprecated - use add_epic_comment with its mentions parameter; this tool will be removed. Post an epic comment with real user mention chips. First call find_epic_comment_mentions and get the user's approval for each selected person. Each choice gives an exact @Name or @@Name token and selected user ID.

add_epic_file

ChatGPT
Attach one file to an epic - a screenshot, a PDF, a small document. The file's bytes are passed base64-encoded, and they pass through the conversation, so this only suits small files, up to 2 MB; for anything larger, or any file on disk, use start_file_upload instead. A file with a name already on the epic is kept under a numbered name rather than replacing it. Set embedInDescription to also show the file inline in the epic's description, appended after whatever is already there - there is no cursor position to place it at, unlike the Project Simple editor, so it always lands at the end. Leave it false to attach the file without touching the description at all. Returns the attached file's id, name, media type and size.

add_epics_to_milestone

ChatGPT
Put epics into a milestone, at the end of its epic order. Use it for "add the payments epic to milestone B" and to fill a milestone you have just created. This is how work gets into a milestone: items come with their epic, not on their own. An epic that is already in another milestone moves out of it, and joining a milestone resets the epic's item planning - its open items become part of the new milestone and work finished earlier is left outside it. Every epic id has to be in this project, or nothing is added and the unknown ones are named. Returns the milestone with the epics now in it.

add_goal

ChatGPT
Add a goal to a project's roadmap. Use it when the user describes something the project is trying to achieve. The three required fields are the three questions Project Simple asks of every goal, so quote the user's own words rather than inventing them - ask if you only have one of the three. The new goal joins the queue as NEXT_IN_LINE; use set_primary_goal to make it the one the project is working towards, which says what it displaces. Returns the new goal in full.

add_idea

ChatGPT
Create a new idea in a project's Ideas list. Use it when the user asks to capture a product idea or feature suggestion; check with them first if the title has to be invented rather than quoted. Title is required; description, size and label are optional and can be set later. Returns the new idea in full, including its reference such as IDEA-42.

add_item

ChatGPT
Create a new work item in a project's backlog. Use it when the user asks for a story, bug or other work item to be added; check with them first if the title has to be invented rather than quoted. A task is not a work item: it is a child of one, so to add a task to an existing item (e.g. "add a QA task to PS-142") use add_item_task or add_item_tasks instead. Title and item type are required; team, epic, sprint, owner, story points, severity and label are optional and can be set later. Required field boundaries Description: the user, customer, team, or business problem and the observable outcome that is needed. Spec: technical, investigative, operational, or design context and instructions for the agent performing the work. Acceptance Criteria: independently verifiable outcomes formatted as checklist items. Description prompt Markdown. Write the item Description in clear, simple English, adapting its structure to the item type and purpose. Describe the problem in the third person from the perspective of the affected user, customer, team, or business: what happened or is missing, who is affected, why it matters, and what outcome is needed. Choose only sections that apply to the item type: - BUG / INCIDENT / SUPPORT_CASE: Summary; Context or Preconditions (when relevant); Steps to Reproduce; Actual Result; Expected Result; User Impact. - STORY / DEV: Problem; Affected User; Current Limitation; Desired Outcome; User Value. - SPIKE: Question or Problem; Why It Matters; Known Context; Unknowns; Decision or Outcome Needed. - DEV_OPS / OPS: Operational Problem; Impact; Current Behavior; Desired Outcome; relevant non-implementation constraints. - TECH_DEBT / DESIGN_DEBT: Current Problem; User or Business Impact; Risk of Leaving It Unresolved; Desired Outcome. Omit sections that do not apply. Do not invent missing facts. Do not use first-person narration and do not use Gherkin unless explicitly requested. Keep Description user-facing: no source code, API payloads, request/response bodies, database details, root-cause assumptions, solution design, or implementation instructions. Put technical context and implementation guidance in Spec. Put independently verifiable completion conditions in Acceptance Criteria. Native references: to link an item, child task, sprint, epic, idea, project, milestone, team or file (or a person in Description), write its visible text and pass it in mentions with its type and id; text without a confirmed choice stays plain. A full URL of this account also becomes a reference without a choice when it names something here: an item of this project (/<account>/<project>-<number>, or a backlog URL with ?i=<project>-<number>), an epic (/<account>/<project>/epic/<number>), a sprint (/<account>/<project>/<team>/sprint/<number>), a milestone (/<account>/<project>/milestones/<title>) or a project (/<account>/projects/<project>); other URLs stay links. Markdown formatting is preserved around references. Read back, each reference shows its text followed by {TYPE:id}, except a task, which shows its text followed by [TASK]; keep that text, with or without the marker, to keep the reference when editing. Never write an id yourself. Existing embedded files are inline Markdown download links with {FILE:id} in their labels; keep the complete link and marker in place to retain the same native file component. Keep Markdown formatting and ordinary link destinations when editing unrelated text. To add a new file reference, use a confirmed FILE choice. A task or file named on the item being written is found by its name; if the item is not known, ask the user which item it belongs to. When the user names a specific resource by its number or name (for example "#124") and its project is not known from context, ask the user which project before writing; otherwise insert a typed reference to it (ITEM for an item) instead of plain text. Spec prompt Markdown. Write the item Spec for the AI agent or specialist performing the work. Adapt the technical structure and level of detail to the item type. Include only relevant implementation, in…

add_item_comment

ChatGPT
Post a comment on a work item as the signed-in user. Use it to record a decision, answer a question on the item or leave a note the team will see - it is visible to everyone who can see the item, so show the user the text before posting anything they did not dictate. Pass replyToCommentId to reply to an existing comment on the same item. For @Name or @@Name, first search with find_item_comment_mentions. If a person is found, confirm the choice and pass it in mentions. If no person is found, post the @ text normally without a mention notification. Returns the comment as it was posted.

add_item_comment_file

ChatGPT
Attach one file to a comment on a work item - not to the item itself; use add_item_file for that. Works on a comment posted by add_item_comment or add_item_task_comment, both being ordinary comments on the item. The file still shows up in list_item_files, which lists the files on the item's comments along with the item's own. The file's bytes are passed base64-encoded, and they pass through the conversation, so this only suits small files, up to 2 MB; a larger file has to be uploaded in Project Simple, or attached to the item, epic or TODO item itself with start_file_upload. Returns the attached file's id, name, media type and size.

add_item_comment_with_mentions

ChatGPT
Deprecated - use add_item_comment with its mentions parameter; this tool will be removed. Post an item comment with real mention chips. First call find_item_comment_mentions and get the user's approval for each selected person. Each choice supplies the exact @Name or @@Name token in text and the selected user ID. Do not call for unconfirmed mentions.

add_item_definition_of_done

ChatGPT
Add a custom Definition of Done entry to an existing work item - the same checkbox a user adds by hand in the item's Definition of Done section. Use it when the user wants an extra completion criterion on an item that no task type tracks, such as "Docs updated" or "Rollout plan approved". The entry starts unticked and is ticked only by a person in the UI; until then it blocks moving the item to DONE. Existing entries are left untouched; this only adds. The name is required, at most 50 characters, and must not repeat an entry already on the item (compared ignoring case); an item converted into an epic cannot take one. There is no tool to tick an entry off: entries that follow tasks are ticked by completing those tasks, the acceptance criteria entry by ticking the checklist in the item's acceptance criteria, and the rest - human-only ones such as 5 Why's included - by a person in the UI. Returns the new entry with its id; read the item again with get_item for the full list.

add_item_file

ChatGPT
Attach one file to a work item. The file's bytes are passed base64-encoded, and they pass through the conversation, so this only suits logs, screenshots and small documents up to 2 MB; for anything larger, or any file on disk, use start_file_upload instead. A file with a name already on the item is kept under a numbered name rather than replacing it. Returns the attached file's id, name, media type and size.

add_item_task

ChatGPT
Add one task to an existing work item of any type - a task is always a child of an item, never a standalone item, so this is the tool whenever the user asks for a task. Use it for a single task; for splitting a whole item into its tasks at once, use add_item_tasks instead - one round trip for the whole breakdown rather than one call per task. Name and type are required; description, spec and estimated hours are optional. Returns the new task.

add_item_task_comment

ChatGPT
Post a comment about one of the item's tasks, as the signed-in user. A task has no comment thread of its own - the comment lands in the item's own thread, same as add_item_comment, tagged with a real reference to the task: it shows as a task chip in Project Simple, the same as if a person had inserted one by hand, and the returned comment's taskId names the task. Show the user the text before posting anything they did not dictate. For @Name or @@Name, search with find_item_comment_mentions first. If a person is found, confirm the choice and pass it in mentions. If nobody is found, keep the @ text as plain text in this comment. Returns the comment as it was posted.

add_item_task_comment_with_mentions

ChatGPT
Deprecated - use add_item_task_comment with its mentions parameter; this tool will be removed. Post an item task comment with its task chip and approved user mention chips. Search targets with find_item_comment_mentions first and get the user's choice. Each choice supplies the exact @Name or @@Name token in text and the selected user ID.

add_item_task_file

ChatGPT
Attach one file to a task - a screenshot, a mockup, a small document. The file's bytes are passed base64-encoded, and they pass through the conversation, so this only suits small files, up to 2 MB; for anything larger, or any file on disk, use start_file_upload instead. A file with a name already on the task is kept under a numbered name rather than replacing it. The file becomes an attachment of the item itself, tagged with this task - use list_item_files to see it afterward and delete_item_file to remove it; there is no task-scoped equivalent of either. Returns the attached file's id, name, media type and size.

add_item_tasks

ChatGPT
Break an existing work item of any type down into several child tasks in one call. Use it for "split this into a design task, a dev task and a QA task" - one round trip for the whole breakdown instead of one add_item_task call per task. Existing tasks on the item are left untouched; this only adds. Each task needs a name and a type; description, spec, estimated hours and confirmed description mentions are optional per task. For @Name or @@Name in a task description, first use find_text_mentions and include that task's selected user IDs in its mentions list. A task's spec holds only what is specific to that task - the item's spec already carries the shared context. Up to 50 tasks per call. Returns the new tasks, in the order given.

add_items_to_epic

ChatGPT
File a batch of work items under an epic in one call. Each item is moved out of whichever epic it was in before, if any; an item already under this epic is still touched and drops to the lowest priority in it - avoid re-adding items that are already there. Use search_items or get_project_overview first to find the items' references. Returns the epic, whose item counts now include the moved items.

add_milestone

ChatGPT
Create a milestone over a date range, optionally against a goal and with epics in it from the start. Use it for "plan a milestone for Q3" once the user has given both dates. A project cannot have a milestone planned ahead of the current one, so this is refused while any milestone is still running or still to come; a new milestone can be created only once the last one has ended or been stopped. Check list_milestones first, and if it is refused tell the user why rather than retrying with other dates. Returns the new milestone with the letter it got and the epics in it, plus movedMilestones: every other milestone whose letter or dates changed. Tell the user about those.

add_retrospective_notes

ChatGPT
Add individual notes to a retrospective, split into What Went Well and What Needs Improvement. Use it after the assistant has synthesized retrospective notes from the conversation, sprint data or user input. Put positive observations in whatWentWell and improvement observations in whatNeedsImprovement. Empty strings are ignored, but at least one real note is required. Returns the notes that were created, including their ids for later edit or delete.

add_todo_item

ChatGPT
Create a TODO item in an account. Use it when the user wants a reminder for themselves or to ask other people to do something; show them the name and description before creating it. Assigned to the current user unless ownerIds says otherwise; every owner has to be a user of the account. Owners other than the caller are emailed. Returns the created item in full.

add_todo_item_comment

ChatGPT
Post a comment on a TODO item as the signed-in user. Its owners and author are notified, so show the user the text before posting anything they did not dictate. Resolve the TODO first with search_todo_items or get_todo_item, and ask which one is meant when more than one matches. Pass replyToCommentId to reply to an existing comment on the same TODO item. For @Name or @@Name, first search with find_todo_item_comment_mentions. If a person is found, confirm the choice and pass it in mentions. If no person is found, post the @ text normally without a mention notification. Returns the comment as it was posted.

add_todo_item_comment_file

ChatGPT
Attach one file to a comment the signed-in user wrote on a TODO item - not to the TODO item itself; use add_todo_item_file for that. The file still shows up in list_todo_item_files and get_todo_item, tagged with the comment's id. The file's bytes are passed base64-encoded, and they pass through the conversation, so this only suits small files, up to 2 MB; a larger file has to be uploaded in Project Simple, or attached to the item, epic or TODO item itself with start_file_upload. Returns the attached file's id, name, media type and size.

add_todo_item_comment_with_mentions

ChatGPT
Deprecated - use add_todo_item_comment with its mentions parameter; this tool will be removed. Post a TODO item comment with real user mention chips; mentioned people are notified. First call find_todo_item_comment_mentions and get the user's approval for each selected person. Each choice gives an exact @Name or @@Name token and selected user ID.

add_todo_item_file

ChatGPT
Attach one file to a TODO item itself - a screenshot, a PDF, a small document. Only the TODO item's owners and its author may add files; to attach a file to a comment, use add_todo_item_comment_file. The file's bytes are passed base64-encoded, and they pass through the conversation, so this only suits small files, up to 2 MB; for anything larger, or any file on disk, use start_file_upload instead. A file with a name already on the TODO item is kept under a numbered name rather than replacing it. Set embedInDescription to also show the file inline in the TODO item's description - for an image the user wants in the description - appended after whatever is already there, since there is no cursor position to place it at. Leave it false to attach the file without touching the description. Returns the attached file's id, name, media type and size.

add_todo_or_idea_comment

ChatGPT
Create a comment on a TODO_ITEM or IDEA. Returns its comment ID. To attach files uploaded in the AI conversation, call attach_conversation_file once per file with targetType=COMMENT and this comment ID.

archive_epic

ChatGPT
Archive an epic, putting it out of the way without changing or unfiling any of its items. Use it when the user wants an epic out of the roadmap or backlog view rather than marked done - archiving does not require its items to be finished. Reversible with restore_epic. Returns the epic with its new status.

assign_item

ChatGPT
Set, change, or clear who owns an item. Passing ownerId replaces the current owner, so use this tool to change an owner as well as to set one. The owner has to be a member of the item's team, as in the UI; get_project_overview lists each team's members with their user ids. To unassign the item, pass unassign=true; ownerId is then ignored. Never send the current owner's id, your own id, or a placeholder value to clear it. Returns the item after the change.

assign_item_task

ChatGPT
Set, or clear, who a task is assigned to. Omit ownerId to unassign the task. Returns the task after the change.

attach_conversation_file

ChatGPT
Copy one uploaded file of any permitted size from the active AI conversation to an ITEM, TASK, EPIC, TODO_ITEM, IDEA or COMMENT without passing its bytes through MCP. Only file IDs supplied with the conversation are accepted; file content is not. The source stays in the conversation. TASK files are embedded in the task description by default. Set embedInDescription=true to put a file in an ITEM, EPIC, TODO_ITEM or IDEA description while preserving existing text. For a ToDo or Idea comment, create the comment with add_todo_or_idea_comment, then use its returned comment ID here with targetType=COMMENT. Existing target permissions apply.

attach_conversation_files

ChatGPT
Copy multiple uploaded files of any permitted size from the active AI conversation to one ITEM, TASK, EPIC, TODO_ITEM, IDEA or COMMENT in a single write without passing their bytes through MCP. Use this when the user wants several files attached to the same target. Pass every requested file ID once; file content is not accepted. Returns one attachment result per file in input order. TASK files are embedded in the task description by default. Set embedInDescription=true to place every file in an ITEM, EPIC, TODO_ITEM or IDEA description. For a ToDo or Idea comment, create the comment first and pass its returned comment ID. Existing permissions apply.

cancel_file_upload

ChatGPT
Discard an upload started with start_file_upload that will not be finished, and the parts already sent. Does nothing when the upload was already completed or discarded. Returns nothing.

cancel_time_off

ChatGPT
Remove a booking of time off. Use it when somebody's plans change. This deletes the entry outright rather than hiding it, and it takes the time off away from everyone it covered, so check who that is with list_time_off first. An entry that also covers people outside this project's teams is refused. Capacity of any sprint the dates fell in goes back up.

change_item_status

ChatGPT
Change the status of one work item. Only statuses a person sets by hand are accepted - NOT_STARTED, RE_OPEN, BLOCKED, OBSOLETE, DONE and UNBLOCKED. UNBLOCKED is only available when the item is BLOCKED and outside an active sprint. Progress statuses, such as IN_DEVELOPMENT or QA Done, are recomputed from the item's tasks whenever the item is saved, so setting them here would not stick. Review statuses require a separate workflow. list_changeable_item_statuses returns the accepted statuses. BLOCKED, UNBLOCKED and OBSOLETE need a message saying why - it is posted on the item and is what the team sees. Returns the item with its new status.

complete_file_upload

ChatGPT
Finish an upload started with start_file_upload, after its uploadScript has sent every part. Attaches the file to the target given at the start; a file with a name already there is kept under a numbered name. Calling it again returns the same file. Fails, naming the first missing part, when not every part has been sent - send the missing parts and call it again. Returns the attached file's id, name, media type and size.

convert_item_to_epic

ChatGPT
Turn a work item into an epic. This is a big change and it is not a rename: it creates an epic from the item, copies the item's files and comments onto it, marks the item OBSOLETE and emails everyone on the project's teams. Confirm with the user before calling it. A JSON backup of the item is saved first, so the whole thing can be undone with undo_item_to_epic_conversion. Returns the new epic, the item as it now stands, and the backup id - report the backup id back to the user, it is how they get out of this.

create_sprint

ChatGPT
Plan a new sprint for one of a project's teams. Use it for "set up the next sprint for the platform team". The sprint is created but not started: fill it with move_items_to_sprint, then call start_sprint. Sprints have no name - they are numbered per team, and the number is assigned here. Dates must not overlap another sprint of the same team. A Marketing team only runs continuous sprints: its sprint is continuous by default and continuous=false is refused. Development, DevOps and UI/UX teams can run either and default to a Scrum sprint. A team has at most one continuous sprint at a time. Returns the sprint with its number, dates, goals and utilization ratio.

create_team_retrospective

ChatGPT
Create or return the retrospective for a team. Use it when the user asks to create a retrospective for a team and optionally invite the team members. Project Simple stores retrospectives under sprints, so pass a sprint id when the user names a specific sprint; otherwise the team's current active sprint is used. If inviteTeamMembers is true, the team's members get the same retrospective request email and TODO items they would get from the UI. Returns the retrospective with its sprint, team, status and notes.

delete_epic_comment

ChatGPT
Remove a comment the signed-in user wrote from an epic. Deleting somebody else's comment is refused, as are the entries Project Simple writes itself - those record what happened to the epic. The comment stops being shown but is not destroyed, so support can still recover it. Returns nothing.

delete_epic_file

ChatGPT
Remove a file from an epic. Any file list_epic_files returns can be removed here, including the ones attached to the epic's comments, but not files on its items. The file stops being shown but is not destroyed, so support can still recover it. Returns nothing.

delete_item_comment

ChatGPT
Remove a comment the signed-in user wrote from a work item. Deleting somebody else's comment is refused, as are the entries Project Simple writes itself - those record what happened to the item. The comment stops being shown but is not destroyed, so support can still recover it. Returns nothing.

delete_item_file

ChatGPT
Remove a file from a work item. Any file list_item_files returns can be removed here, including the ones attached to the item's tasks or comments. The file stops being shown but is not destroyed, so support can still recover it, and the removal is recorded on the item. Returns nothing.

delete_item_task

ChatGPT
Remove a task from a work item. Unlike an item or an epic, a task can genuinely be deleted through MCP - it is a soft delete with a real undo: the returned restoreToken can be passed to restore_item_task to bring the task back exactly as it was, including its id. Losing the token means the task is gone for good, so show or keep it if the user might want it back. Returns the deleted task's id and its restore token.

delete_todo_item

ChatGPT
Delete a TODO item that no longer needs doing. Only the item's owners and its author may delete it, the same rule as the UI. The item disappears from every list and from search; confirm with the user first, since no tool restores it. To keep a record of work that was done, check the item off instead.

delete_todo_item_comment

ChatGPT
Remove a comment the signed-in user wrote from a TODO item. Deleting somebody else's comment is refused, as are the entries Project Simple writes itself. Confirm with the user first. The comment stops being shown but is not destroyed, so support can still recover it. Returns nothing.

delete_todo_item_file

ChatGPT
Remove a file from a TODO item. Only the TODO item's owners and its author may remove files. Any file list_todo_item_files returns can be removed here, including the ones on the TODO item's comments. Confirm with the user first. The file stops being shown but is not destroyed, so support can still recover it. Returns nothing.

edit_epic_comment

ChatGPT
Replace the text of a comment the signed-in user wrote on an epic. Editing somebody else's comment is refused, as are the entries Project Simple writes itself. The new text replaces the old one entirely. For @Name or @@Name, search with find_epic_comment_mentions first. If a person is found, confirm the choice and pass it in mentions. If nobody is found, keep the @ text as plain text in this edit. Returns the comment after the edit.

edit_epic_comment_with_mentions

ChatGPT
Deprecated - use edit_epic_comment with its mentions parameter; this tool will be removed. Replace your epic comment with text and approved real user mention chips. Search and confirm each target with find_epic_comment_mentions first.

edit_item_comment

ChatGPT
Replace the text of a comment the signed-in user wrote. Editing somebody else's comment is refused, as are the entries Project Simple writes itself. The new text replaces the old one entirely. For @Name or @@Name, search with find_item_comment_mentions first. If a person is found, confirm the choice and pass it in mentions. If nobody is found, keep the @ text as plain text in this edit. Returns the comment after the edit.

edit_item_comment_with_mentions

ChatGPT
Deprecated - use edit_item_comment with its mentions parameter; this tool will be removed. Replace your item comment with text and approved real user mention chips. Search and confirm each target with find_item_comment_mentions first. The replacement text must include each selected token.

edit_todo_item_comment

ChatGPT
Replace the text of a comment the signed-in user wrote on a TODO item. Editing somebody else's comment is refused, as are the entries Project Simple writes itself, and so is a comment that is not on the named TODO item. The new text replaces the old one entirely. For @Name or @@Name, search with find_todo_item_comment_mentions first. If a person is found, confirm the choice and pass it in mentions. If nobody is found, keep the @ text as plain text in this edit. Returns the comment after the edit.

edit_todo_item_comment_with_mentions

ChatGPT
Deprecated - use edit_todo_item_comment with its mentions parameter; this tool will be removed. Replace your TODO item comment with text and approved real user mention chips. Search and confirm each target with find_todo_item_comment_mentions first.

estimate_sprint_capacity

ChatGPT
Work out what a team could deliver over a window, before any sprint exists for it. Use it when planning the next sprint - to choose its dates, or to see what a different utilization ratio would buy. Time off already in the calendar for that team over those dates is taken into account, so the answer matches what the real sprint would get. Planned effort comes back as zero: there is nothing in it yet. Returns the same figures as get_sprint_capacity.

find_epic_comment_mentions

ChatGPT
Search people in the epic's project before posting. Both @ and @@ return individual users because an epic has no single team. Show matches to the user and get their choice before writing.

find_item_comment_mentions

ChatGPT
Search people before posting. @ searches people on the item's team, or in its project when it has no team; @@ searches people in the item's project. Show matches to the user and get their choice before writing. Returns the selected USER ID.

find_text_mentions

ChatGPT
Find a person by first or last name before writing @Name or @@Name in a comment or description. @ searches people in the specified team, falling back to the project or account when no team or project is available. @@ searches people in the project, or across the account when there is no project. Show matches and let the user select a person before writing a real mention with the selected USER ID. If no person matches, save the original @ text as plain text, without a mention choice.

find_todo_item_comment_mentions

ChatGPT
Search people to mention in a TODO item comment before posting. A TODO item has no team, so both @ and @@ search the users of the account. Show matches to the user and get their choice before writing.

get_ai_activity

ChatGPT
Summarise an account's AI usage, as the AI Activity Log page shows it. Use it for "how much is AI used", "which features or projects use AI most", "who uses it", "how many requests failed" and trends over time. Account administrators only. Covers all time, as the page does with no range selected, unless from is given; a period with from covers at most a year. Pass from and to for questions about a period such as this week or last month. Returns total, failed and success rate, average duration, feedback counts, requests per type (GENERAL is the AI Assistant), the change against the period before, requests per day, and the projects and users with the most requests.

get_current_sprint

ChatGPT
Get the sprint each team on a project is running right now, or is about to start. Use it for "what is in the current sprint" and as the source of the sprint id for search_items. Returns one entry per team - a project worked on by several teams has several current sprints - with the sprint id, its number, team, status, dates, goals and item counts. Returns an empty list when no team on the project has an open sprint.

get_epic_milestone_items

ChatGPT
Show one epic's items split by where they stand in its milestone: planned, excluded, or completed within the milestone. Use it before set_epic_milestone_items, and to answer "what of this epic is actually in the milestone" - an epic in a milestone does not necessarily commit everything inside it. Planned and completed items count towards the milestone; excluded ones do not. Items finished before the epic joined are outside the milestone and are not listed. Returns the three groups, plus the milestone the epic is in.

get_goal

ChatGPT
Get one roadmap goal with the milestones it is being delivered through. Use it after list_goals when the user asks about a specific goal, or to see the dates the goal is committed to. Returns the goal's text, status and roll-up counts, plus each of its milestones with its dates and progress. The goal's counts cover the epics under its features; the milestone's counts cover every epic in the milestone, so the two can differ.

get_idea

ChatGPT
Read one idea in full. Use it after search_ideas or when the user names an idea reference such as IDEA-42, a bare idea number, or an idea UUID. Returns the idea summary plus the plain-text description, the caller's own vote, and resolved titles and sequence numbers for converted backlog items or epics.

get_item

ChatGPT
Read one work item in full. Use it after search_items, or whenever the user names an item such as PS-142. Returns the item summary plus the description and acceptance criteria as Markdown (long descriptions are truncated), the spec in full, the tasks in their order with their order, owners, estimates, completion, created date, the team type the task came from, read-only (task type not supported by the item's current team type), outdated-estimate and checklist-fully-checked flags and their own specs, and every Definition of Done entry in full: its id, name, the task types and item status it tracks, whether a person added it by hand (manual), how it gets ticked (completion: TASKS by finishing the item's tasks of those types, ACCEPTANCE_CRITERIA by the checklist in the acceptance criteria, MANUAL only by a person in the UI), whether it is human-only such as 5 Why's, and whether it is ticked already. No tool ticks an entry; add_item_definition_of_done adds one. Also returns ready, complexity, sprint priority, story points modified date, created and modified date and user, last modified source, last status change in sprint, the not-estimated-task, demo video, post mortem, active story poker, read-only and converted-to-epic flags, and statusChange: why the item is BLOCKED, UNBLOCKED, OBSOLETE or in PM or client review - the reason, its author and date, and the status before and after. Read it before implementing the item or any of its tasks.

get_item_analytics

ChatGPT
Get the delivery analytics Project Simple keeps for one work item. Use it for "why is PS-142 taking so long", "where is the time going on this bug" and for anything about lead or cycle time on a single item. Returns when the item's clock started, the total time its tasks were worked on split by task type, how long it sat in each status it passed through and whether the team changed, and the work that appeared or disappeared after it was planned. All durations are milliseconds; convert them before showing them to anyone.

get_milestone

ChatGPT
Get one milestone with the epics that make it up and the goal it is delivering. Use it to answer "what is in milestone B" and to check the scope before changing it. Returns the milestone's dates, derived status and progress, the epics in it, and the goal attached to it if there is one. Which of an epic's items are inside the milestone's scope is a separate question - get_epic_milestone_items answers it.

get_my_work

ChatGPT
Get the current user's own work in an account. Use it for "what am I working on", "what is on my plate" and standup-style questions. Returns three groups - Active Work (what the user is set to work on now), Worked On (items they have a task on but do not own) and Completed - plus counts of items, of the user's still open tasks and of the hours those tasks are estimated at.

get_project_overview

ChatGPT
Summarise the current state of one project. Use it to answer "how is project X doing" before drilling into items or sprints. Returns the project, the teams working on it with their members (user id, name and email - use these ids for owners and assignees), the sprint each team is running now, item counts split into total, not completed, in work and completed, and the project's 5 Why's severity rule.

get_retrospective

ChatGPT
Read one retrospective in full. Use it before changing or deleting notes, or when a user asks what has already been captured. Find it by its own id, or by the sprint it belongs to - pass sprintId from get_current_sprint, list_sprints or get_sprint (get_sprint's hasRetrospective says whether there is one). One of the two is enough; when both are passed they must name the same retrospective. Returns the sprint and team it belongs to, the status, any AI insights summary already stored on it, and every individual note with its id and whether it is in What Went Well or What Needs Improvement. Until the retrospective is REVEALED or COMPLETED, notes stay private as they do in the UI: only the signed-in user's own notes come back. The retrospective id it returns is the one the other retrospective tools take.

get_sprint

ChatGPT
Read one sprint in full, including the settings a person edits. Use it before update_sprint - that tool changes only the fields it is given, so it needs to know what the others are - and whenever the summary from list_sprints is not enough. Historical sprints remain readable after their team is removed from the project; teamRemovedFromProject marks that case. Returns everything list_sprints returns plus the utilization ratio, the work-in-progress limit and cycle length of a continuous sprint, the review status, the sprint summary, the halt reason if it was halted, and the planned effort in hours of everything in it.

get_sprint_analytics

ChatGPT
Get the Sprint Analytics report for one sprint over a window of 1, 3 or 6 sprints. Use it for "how is the team performing", "how did this sprint go", velocity, cycle or lead time, Blocked time, quality, health check and team insights, and to check figures against the Sprint Analytics page. Use get_item_analytics for one item instead. Returns the project, the anchor sprint, the sprints the window actually covers, and the metrics grouped by section. Each metric has a stable id, a value, a unit, a scope (PROJECT or TEAM), whether the window applies to it, and its availability. A value of 0 is measured; a null value with availability NO_DATA or INSUFFICIENT_HISTORY means there is no figure, so never report it as 0. Cycle Time by Status keeps Blocked even though the page's chart hides it, and blockedCycleTime under PERFORMANCE is the separate Blocked Cycle Time metric. The 1, 3 and 6 sprint windows are alternatives, so never add their values together. Refused for a sprint that has not started.

get_sprint_capacity

ChatGPT
Work out how many hours a sprint's team can deliver in it, and how much is already committed. Use it for "can we fit this in" and "is the sprint over-committed". Capacity is eight hours per engineer per weekday, minus the days they are marked off in the calendar, times the sprint's utilization ratio - the share of the day planned as delivery time. Returns that ratio alongside the hours, because the two only mean something together, plus work days, engineer-days, days lost to time off, the planned effort of the items in the sprint, what is left and the committed percentage. A negative remaining figure means the sprint is over-committed.

get_sprint_readiness

ChatGPT
Check whether a sprint can be started, and if not, say exactly what is in the way. Use it before start_sprint, and to answer "why can't we start the sprint". Project Simple requires every item in a sprint to be broken down into tasks and estimated in hours - and, unless the sprint is continuous, in story points - and allows only one running sprint per team. Returns whether it can start, one line per unmet rule, and the items behind each: those with no tasks, no hour estimate and no story points.

get_story_poker_history

ChatGPT
Show an item's Story Poker sessions, newest first: status (open, completed, or abandoned when replaced before completion), who started each and when, who voted with what, the average of the votes and the card it rounds up to (what the Story Poker screen suggested). The number a session was completed with, and whether that was the suggestion or a hand-picked card, is not recorded. Votes in a session that is still open are shown without their numbers, except the caller's own, and it has no average. Also returns the item's current story points.

get_team

ChatGPT
Read one team as Team Settings shows it. Use it for "who is on this team", "which roles are we missing", "which projects does this team work on" and "how does our sprint setup affect planning". Returns the members sorted by name with their user id, email, team role, whether they manage the team and whether they count towards sprint capacity (engineers do); the projects the team works on with their sprint length in weeks and utilization ratio; the sprint setup - SCRUM or CONTINUOUS, the active sprint with its dates, WIP limit, continuous cycle length and utilization ratio; and the task types the team's items can be broken down into. Visible to the team's members and to anyone who can read one of its projects.

get_time_off_summary

ChatGPT
Per-person totals of time off over a window - how many bookings, and how many days off, clipped to the window's edges. Use it for "how much time off has the team booked this quarter" or "who's taken the most time off" - for the individual bookings behind those totals use list_time_off instead. Narrowed the same way list_time_off is: to one team, to an explicit set of people, or - when neither is given - the whole project. The window may be at most 400 days.

get_todo_item

ChatGPT
Read one TODO item in full. Use it after search_todo_items or when the user names a TODO reference such as TODO-42, a bare number, or a TODO item UUID. Returns the item summary - name, status, owners and who has checked it off, completion rule, deadline, related resource - plus the plain-text description and the attached files with download links, including the files on its comments tagged with commentId. Read the discussion with list_todo_item_comments.

get_todo_item_metrics

ChatGPT
Count the current user's open TODO items in an account. Use it for "what is on my plate", "anything overdue?" and morning-brief style questions before listing the items themselves. Returns how many active items the user still has to do, how many active items they created, and of their own how many are urgent, due today and past due.

halt_sprint

ChatGPT
Stop a running sprint before its end date. Use it when a sprint is being abandoned - a change of priorities, an incident, a team being pulled elsewhere. This is not a pause and there is no un-halt: the sprint ends there, its unfinished items are carried over to the next sprint (created if there is none), a sprint review is opened and a snapshot of the team is taken. Say all of that back to the user before calling it. Returns the halted sprint.

invite_retrospective_team_members

ChatGPT
Invite or re-invite the team members to complete a retrospective. Use it when the retrospective already exists and the user asks to send the request to the team. This follows the same flow as the UI: members receive the retrospective request email and TODO items are refreshed for the retrospective. Returns the retrospective after the request has been made.

invite_to_story_poker

ChatGPT
Open a Story Poker session on an item and ask a team - or chosen members of it - to vote on its story points. Each invitee gets a Story Poker ToDo with a link to the voting screen. Only members of the team can vote, so every invitee must belong to it; omit invitees to invite the whole team. A team with no members is refused. If the item already has a session, it is replaced: its outstanding ToDos are withdrawn and its votes stay in get_story_poker_history. Confirm with the user before replacing a session that is still open. Returns the new session and who was invited.

list_changeable_item_statuses

ChatGPT
List the statuses change_item_status will accept. Use it before changing a status if the user named one that might be task-driven, rather than guessing and getting an error. Returns status names. UNBLOCKED is only available for a BLOCKED item outside an active sprint. Other statuses are recomputed from tasks and definition of done, or require a separate review workflow.

list_epic_comments

ChatGPT
Read the comments on an epic, newest first. Use it whenever the question is about why something happened to the epic - why it was archived, what was decided about its scope. By default only what people wrote is returned; pass includeAudit=true to also get the entries Project Simple writes itself, which come back flagged as audit with no text of their own. Returns author, date, text, attachments and whether the comment is a reply.

list_epic_files

ChatGPT
List the files attached to an epic, most recent first. Use it to answer what is attached, or to find the id of a file to remove. Files attached to the epic's comments are included, alongside the epic's own; files on the epic's items are not. Returns each file's id, name, media type and size - not its contents.

list_epics

ChatGPT
List a project's epics - the groups of work a milestone is built out of. Use it to find the epic ids for add_epics_to_milestone, and to answer "what epics are not in a milestone yet" by passing inMilestone=false. Returns each epic's id, number, name, status, progress, label and teams. Archived epics are left out unless asked for.

list_goals

ChatGPT
List a project's roadmap goals, the primary one first and the rest in roadmap order. Use it for "what are we working towards", to find the id of a goal, and before set_primary_goal so the user can be told which goal is primary today. Returns each goal's text, its status (CURRENT means primary, then NEXT_IN_LINE, FUTURE and CLOSE), the milestone it is delivered through, and how many epics and items sit under it. Those counts cover the epics under the goal's features only; epics that are merely in the goal's milestone are counted by get_milestone, not here.

list_item_comments

ChatGPT
Read the comments on a work item, newest first. Use it whenever the question is about why something happened - why an item is blocked, what a reviewer asked for, what was agreed. By default only what people wrote is returned; pass includeAudit=true to also get the entries Project Simple writes itself for status changes and edits, which come back flagged as audit with no text of their own. Returns author, date, text, attachments and whether the comment is a reply.

list_item_conversion_backups

ChatGPT
List the backups saved when a work item was converted into an epic, newest first. Use it to find the backup id for undo_item_to_epic_conversion when the user did not keep it, or to answer whether a conversion can still be undone. Returns each backup's id, when and by whom the conversion ran, the epic it produced, and whether the backup has already been used.

list_item_files

ChatGPT
List the files attached to a work item, most recent first. Use it to answer what is attached, or to find the id of a file to remove. Files attached to the item's tasks and to its comments are included, alongside the item's own. Returns each file's id, name, media type and size - not its contents.

list_milestone_items

ChatGPT
List the work items in a milestone, across every epic in it. Use it for "what is left in milestone B" and to see the milestone as a list of items rather than of epics. Items excluded from the milestone, and work an epic had already finished before it joined, are not in the milestone and are left out. Returns a page of items with their reference, title, status, owner, epic and sprint, and the total so you can tell whether you have seen all of them.

list_milestones

ChatGPT
List a project's milestones in date order, earliest first. Use it for "what is coming up", to find the id of a milestone, and as the source of the milestone id the other milestone tools take. Milestones are lettered A, B, C by date rather than named, so quote the letter with the dates and never on its own. Returns each milestone's letter, derived status, dates, days left, epic count and item counts. Milestones that have ended or were stopped are left out unless asked for.

list_my_accounts_and_projects

ChatGPT
List the accounts the current user belongs to and the projects they can read in each. Call this first: every other Project Simple tool needs a project id or an account id, and this is where those ids come from. Returns an account id and name per account, and per project its id, name, key prefix (the PS in PS-142), whether it is a demo project, and its 5 Why's severity rule.

list_sprints

ChatGPT
List one page of sprints from every team currently or formerly working on a project, most recent first. Use it to look back over past sprints or to find the id of a sprint that is not the current one. Includes completed and halted sprints by default. Each sprint has teamRemovedFromProject to distinguish historical teams from current teams. Returns sprint id, number, team, status, dates and goals, without item counts, plus totalCount and statusCounts for all pages and hasMore. To list all sprints, start at page 0 and request the next page while hasMore is true.

list_teams

ChatGPT
List the current user's teams in an account, with the projects each one works on. Use it on account-level pages such as My Work or Calendar, or whenever a question spans "my teams" and no project is selected: it is where the team and project ids for get_team, get_current_sprint, get_sprint_capacity and list_time_off come from. Returns one page of teams with their id, name, type, the user's role in each, and the projects (id, name, prefix) the user can read, plus the total number of teams.

list_time_off

ChatGPT
List the time off booked for a project's teams over a window. Use it for "who is off during the next sprint", before promising a date, or to look up someone's time off history. One entry can cover several people - a team offsite is one booking, not five. Returns each entry's id, the people it covers, its description and its first and last day. The window may be at most 400 days.

list_todo_item_comments

ChatGPT
Read the comments on a TODO item, newest first. Use it when the user asks what was said about a TODO, or to find the id of a comment to reply to, edit or delete. By default only what people wrote is returned; pass includeAudit=true to also get the entries Project Simple writes itself. Returns id, author, date, text, attachments and the comment it replies to.

list_todo_item_files

ChatGPT
List the files on a TODO item, most recent first - the same files the TODO's File tab shows. Use it to answer what is attached, to get a download link, or to find the id of a file to remove. Files attached to the TODO item's comments are included alongside its own, each once, with commentId naming the comment it came from; commentId is empty for a file on the TODO item itself. Returns each file's id, name, media type, size and a short-lived download link - not its contents.

move_item_to_epic

ChatGPT
File a work item under an epic, or take it out of the one it is in. Use it when the user wants an item grouped with an epic's other work. Omit the epic id to remove the item from its current epic; the item itself is not affected either way. Returns the item with its new epic.

move_item_to_sprint

ChatGPT
Move a work item into a sprint, or back to the backlog. Use it for "pull PS-142 into this sprint" and "push PS-142 back to the backlog". Omit the sprint id to move the item to the backlog. Adding an item to a sprint that is already running is refused unless the item is estimated and broken down into tasks, which is Project Simple's rule, not this tool's. When removing an item from an active sprint, a nonblank message is required if its status is not NOT_STARTED, BLOCKED, OBSOLETE or DONE. The message becomes an item comment, and the removal changes that item's status to BLOCKED. Items in those four statuses keep their status. Check the returned status and report any change. Returns the item with its new sprint and status.

move_item_to_team

ChatGPT
Hand a work item to a team, or take it off the team it is on. Use it for "give PS-142 to the platform team". Omit the team id to leave the item without a team. This moves more than the team field: the item leaves the sprint of the team it is leaving, and anyone assigned to it who is not on the new team is unassigned. Say so when reporting back. Returns the item with its new team.

move_items_to_sprint

ChatGPT
Move several work items into a sprint, or push several back to the backlog. Use it for sprint planning - "pull PS-142, PS-143 and PS-150 into the next sprint" - where move_item_to_sprint would mean one call each. Omit the sprint id to move them all to the backlog. Each item is moved on its own: an item Project Simple refuses - adding an unestimated or unbroken-down item to a running sprint is refused, which is its rule, not this tool's - leaves the others where this call put them. When removing items from an active sprint, a nonblank message is required for each item whose status is not NOT_STARTED, BLOCKED, OBSOLETE or DONE. It becomes a comment on each such item, and each changes to BLOCKED. Items in those four statuses keep their status. Check and report each returned status. Returns one line per item saying whether it moved and, if not, why.

promote_idea_to_backlog_item

ChatGPT
Convert an idea into a new backlog item. Use it when the user decides an idea should become planned work. The existing service creates the item, marks the idea CONVERTED, stores the created item id on the idea, and migrates the idea's comments to the new item. Returns the updated idea.

promote_idea_to_epic

ChatGPT
Convert an idea into a new epic. Use it when the user decides the idea is a larger body of work. The existing service creates the epic, marks the idea CONVERTED, stores the created epic id on the idea, and migrates the idea's comments to the new epic. Returns the updated idea.

promote_idea_to_initiative

ChatGPT
Convert an idea into a new Initiative of a roadmap Goal. Use it when the user accepts an idea as roadmap-level work under a Goal. Find the Goal with list_goals; a closed Goal is refused. The name and change default to the idea's title and description text; goalReason, why the Initiative matters to the Goal, has no default. Limits: name 46, change 140, goalReason 190 characters - shorten a long title or description rather than truncating meaning. The Initiative is added as the Goal's last one, the idea is marked CONVERTED and keeps a link to it. Returns the updated idea.

promote_idea_to_roadmap_feature

ChatGPT
Convert an idea into a new roadmap feature. Use it when the user accepts an idea as roadmap-level functionality rather than a backlog item or epic. The existing service creates the feature, marks the idea CONVERTED, and stores the created feature id on the idea. Returns the updated idea.

purge_epic

ChatGPT
Delete an epic outright, with the comments and attachments on it. Super administrators only; everyone else gets a refusal. This is for an epic that should never have existed - usually one an accidental conversion created - and it cannot be undone. Items in the epic are unlinked, never deleted, and if the epic came from converting an item, that item is restored from its backup first and keeps the comments and files it never gave up. Ask the user to confirm, then call with confirm=true; without it the tool refuses. Returns what was removed and the restored item, if there was one.

remove_epics_from_milestone

ChatGPT
Take epics out of a milestone, leaving the epics and their items untouched. Use it for "that epic is not part of milestone B after all". The epic keeps all its work and simply stops counting towards the milestone; every item in it becomes planned again, so putting it back is not free. Every epic has to be in this milestone, or nothing is removed - an epic that sits in a different milestone would otherwise be taken out of that one instead. Returns the milestone with what is left in it.

remove_items_from_epic

ChatGPT
Take a batch of work items out of an epic, back to no epic. Every item given must currently be under this epic - the whole call is refused if even one is not, rather than moving the rest and silently skipping it. Returns the epic, whose item counts now exclude the removed items.

resolve_todo_item

ChatGPT
Check a TODO item off for the current user. Use it when the user says they have done it. Only an owner can check it off. The item itself becomes COMPLETED once its completion rule is met: right away under ANYONE, or when the last owner checks it off under EVERYONE. Returns the item with the owners' updated completion.

restore_epic

ChatGPT
Bring an archived epic back into the roadmap and backlog views. Its items and their statuses were never touched by archiving, so nothing about them changes here either. Returns the epic with its new status.

restore_item_task

ChatGPT
Undo a delete_item_task, using the restoreToken it returned. The task comes back exactly as it was at the moment it was deleted, including its original id. Returns the restored task.

schedule_time_off

ChatGPT
Mark one or more people off between two dates. Use it for "Sam is on holiday the first week of March". Both dates are inclusive and the whole of each day is taken off; there is no half day. Everyone named has to be on a team that works on this project. This changes the capacity of every sprint the dates fall in, straight away - say so when the dates land inside a running sprint. Returns the entry that was created, with the id needed to cancel it.

search_ideas

ChatGPT
Search a project's Ideas list by text and by attributes. Use it when the user asks to find ideas, review the ideas board, or locate an idea before editing, voting or promoting it. Returns a page of ideas with ids, references such as IDEA-42, status and readiness flags, vote counts, conversion ids, size, label and timestamps.

search_items

ChatGPT
Search the work items of one project by text and by attributes, or search its backlog. Use it to find items to talk about; use get_item afterwards when the full description or the task breakdown is needed. To check whether one named item such as PS-142 exists, call get_item instead: it says when there is no such item, where an empty search page cannot tell missing from filtered out. For every item in the project - in a sprint or not, open or closed - leave scope at ANY and pass onlyOpen=false; with onlyOpen omitted, ANY returns open items only. To find something in the backlog pass scope=BACKLOG: that covers open items outside a sprint plus done and obsolete items, returned in backlog order (prioritized, then unprioritized, then closed items newest first), and leaves out items that are in a sprint. Pass onlyOpen=true for the active backlog only, or PRIORITIZED_BACKLOG / UNPRIORITIZED_BACKLOG for one part of it. Read totalCount and page through pageSize=100 pages until every item is listed before calling a list complete. Returns a page of items - id, reference such as PS-142, a url that opens the item, title, type, status, owner, sprint, epic, story points, estimate, label, backlog priority and completion time - plus the total number of matches.

search_todo_items

ChatGPT
Search an account's TODO items by text and by attributes. Use it when the user asks what they have to do, what they asked others to do, what is overdue, or to find a TODO before editing or checking it off. TODO items belong to an account, not a project, and each has one or more owners who check it off individually. Returns a page of items with ids, references such as TODO-42, status, owners and their completion, deadline, urgency and label, open items first and the most urgent at the top.

set_epic_milestone_items

ChatGPT
Choose which of an epic's items are part of its milestone. Use it for "that work is not part of this milestone" (excluded=true) and to put items back into scope (excluded=false). This changes what the milestone is measured against, not the items themselves - nothing is moved, closed or deleted. The epic has to be in a milestone, every item has to be in that epic, and items already finished cannot be moved between the groups; the call is refused rather than partly applied. Returns the epic's items in their groups after the change.

set_item_demo_video

ChatGPT
Make a video the work item's demo video, the one the item shows by default. Use it whenever the user asks for a demo video: first attach the file to the item - attach_conversation_file, add_item_file or start_file_upload - then call this with the file id it returned; or pass the link of a video hosted elsewhere (Loom, Drive, YouTube, ...). Exactly one of fileId and videoUrl. The demo video shown before is replaced as the default but stays in the item's list of demo videos. Returns the demo video's id, its file id or link, and whether it is current.

set_milestone_dates

ChatGPT
Move a milestone's start and end dates. Use it for "push milestone B out by two weeks". Send both dates, not only the one that changes. Milestones cannot overlap, so a later end date pushes back every milestone it would have run into, by the same number of days, one after another - a two-week slip moves the rest of the roadmap by two weeks. A milestone that has already ended cannot be moved, and one that is running now keeps the start date it began with. Returns the milestone with its new dates, plus movedMilestones: every other milestone the change pushed back, with its new dates. Tell the user about those.

set_primary_goal

ChatGPT
Make one goal the project's primary goal - the one the roadmap leads with and the teams are working towards. Use it for "we are focusing on X now" and "move X to the top of the roadmap". A project has one primary goal, so this also moves whichever goal was primary back to next in line; both goals come back in the result and are worth naming when you report. Neither goal's milestone is touched: stopping a milestone files a review and empties it of unfinished epics, which is a decision about the plan rather than about the order of the roadmap, and stays in Project Simple. To undo, run the tool again on the goal that was displaced.

start_file_upload

ChatGPT
Start uploading a file that is on the user's machine to a work item, a task, an epic, an idea, a goal or a TODO item. Use it for any file larger than about 1 MB, or whenever the file is on disk - videos, large PDFs, archives - instead of add_item_file and the other base64 tools; any size the account's plan allows works (Team 500 MB, Business 2 GB, Enterprise 5 GB). Never read the file into the conversation. Run the returned uploadScript in a shell after setting FILE to the file's path: it sends the file in parts with curl. Then call complete_file_upload with the uploadToken. If the user cannot run shell commands, the file has to be uploaded in Project Simple instead. Refused before anything is sent when the file is over the plan's limit. Returns the uploadToken, part size and count, the part URL template and the uploadScript.

start_item

ChatGPT
Start work on one item in the team's active sprint. Use it right before implementing an item, or when a user says "start PS-142". This is the equivalent of the UI's "Start work": it transitions the item through IN_PROGRESS to stamp the cycle-time start and the estimation baseline, and then Project Simple recomputes the visible status (IN_DEVELOPMENT, IN_QA, ...) from the task breakdown on save. Refused if the item is not NOT_STARTED / RE_OPEN, is not in an active sprint, or has no task breakdown / definition of done. Returns the item with its updated status.

start_sprint

ChatGPT
Start a sprint that has been planned and filled. Project Simple refuses to start a sprint whose items are not all broken down into tasks and estimated, refuses an empty sprint, and refuses a second sprint on a team that already has one running. Those rules hold here - this tool cannot waive them. Call get_sprint_readiness first if you want to know what is missing before trying; a refusal here names the items in the way. Returns the sprint once it is running.

submit_story_poker_vote

ChatGPT
Cast or change the signed-in user's Story Poker vote on an item's open session. points must be a card: 0, 1, 2, 3, 5, 8, 13 or 21. Voting again replaces the earlier vote. Only members of the session's team may vote, and only while the session is open. Returns the session; other people's numbers stay hidden until it is completed, but who has voted is shown.

undo_item_to_epic_conversion

ChatGPT
Undo a conversion of a work item into an epic. Restores the item from the backup taken before the conversion - its title, description, status and epic - and deletes the epic the conversion created. The item's own comments and files never left it, so they need no restoring; the copies of them that were put on the epic go away with the epic. A backup can only be spent once and the epic goes for good, so this is as far-reaching as the conversion was. Confirm with the user before calling it, naming the item and the epic. Returns the restored item and what the undo actually did.

unresolve_todo_item

ChatGPT
Reopen a TODO item for the current user after it was checked off by mistake. Only an owner can reopen it, and an item Project Simple generated (a review request, a timesheet reminder) cannot be reopened at all. A COMPLETED item goes back to ACTIVE when the reopened owner means its completion rule is no longer met. Returns the item with the owners' updated completion.

update_epic

ChatGPT
Change the content of an existing epic - its name, description, size or label. Send only the fields that should change; anything omitted is left exactly as it is, so there is no need to read the epic first and repeat the rest. Send an empty string to clear the description or the label. Use archive_epic and restore_epic for the epic's status, and move_item_to_epic to change which items it holds - neither can be set here. Returns the epic in full after the change.

update_goal

ChatGPT
Change the wording of an existing roadmap goal - what it is, why it matters, its timeline, why now, or its video link. Send only the fields that should change; anything omitted is left exactly as it is, so there is no need to read the goal first and repeat the rest. Send an empty string to clear why-now or the video link; the first three cannot be blanked. This tool cannot change which goal is primary - that is set_primary_goal - and cannot attach or detach a milestone. Returns the goal in full after the change.

update_idea

ChatGPT
Change an existing idea's content or board attributes: title, description, label, size, readiness or status. Send only the fields that should change; omitted fields remain unchanged. Send an empty string to clear the description or label. Use vote_on_idea for voting and promote_idea_to_* tools for backlog item, epic or roadmap feature conversion. Returns the idea in full after the change.

update_item

ChatGPT
Change the content of an existing work item. Send only the fields the user asked to change; anything omitted is left exactly as it is, so there is no need to read the item first and repeat the rest. An empty string changes nothing, and neither does 0 for story points or complexity the item does not have yet. To remove a value, name the field in clearFields - only when the user explicitly asks to remove it. A description, acceptanceCriteria or spec you send replaces the whole field. When the user asks to add something to one - a sentence, a reference - send its current text with the addition and every other word unchanged; on an empty field send only the addition. Do not rewrite the rest or fill in sections nobody asked for. Use change_item_status for the status, assign_item for the owner and the move_item_to_* tools for epic, sprint and team - none of those can be set here. Required field boundaries Description: the user, customer, team, or business problem and the observable outcome that is needed. Spec: technical, investigative, operational, or design context and instructions for the agent performing the work. Acceptance Criteria: independently verifiable outcomes formatted as checklist items. Description prompt Markdown. Write the item Description in clear, simple English, adapting its structure to the item type and purpose. Describe the problem in the third person from the perspective of the affected user, customer, team, or business: what happened or is missing, who is affected, why it matters, and what outcome is needed. Choose only sections that apply to the item type: - BUG / INCIDENT / SUPPORT_CASE: Summary; Context or Preconditions (when relevant); Steps to Reproduce; Actual Result; Expected Result; User Impact. - STORY / DEV: Problem; Affected User; Current Limitation; Desired Outcome; User Value. - SPIKE: Question or Problem; Why It Matters; Known Context; Unknowns; Decision or Outcome Needed. - DEV_OPS / OPS: Operational Problem; Impact; Current Behavior; Desired Outcome; relevant non-implementation constraints. - TECH_DEBT / DESIGN_DEBT: Current Problem; User or Business Impact; Risk of Leaving It Unresolved; Desired Outcome. Omit sections that do not apply. Do not invent missing facts. Do not use first-person narration and do not use Gherkin unless explicitly requested. Keep Description user-facing: no source code, API payloads, request/response bodies, database details, root-cause assumptions, solution design, or implementation instructions. Put technical context and implementation guidance in Spec. Put independently verifiable completion conditions in Acceptance Criteria. Native references: to link an item, child task, sprint, epic, idea, project, milestone, team or file (or a person in Description), write its visible text and pass it in mentions with its type and id; text without a confirmed choice stays plain. A full URL of this account also becomes a reference without a choice when it names something here: an item of this project (/<account>/<project>-<number>, or a backlog URL with ?i=<project>-<number>), an epic (/<account>/<project>/epic/<number>), a sprint (/<account>/<project>/<team>/sprint/<number>), a milestone (/<account>/<project>/milestones/<title>) or a project (/<account>/projects/<project>); other URLs stay links. Markdown formatting is preserved around references. Read back, each reference shows its text followed by {TYPE:id}, except a task, which shows its text followed by [TASK]; keep that text, with or without the marker, to keep the reference when editing. Never write an id yourself. Existing embedded files are inline Markdown download links with {FILE:id} in their labels; keep the complete link and marker in place to retain the same native file component. Keep Markdown formatting and ordinary link destinations when editing unrelated text. To add a new file reference, use a confirmed FILE choice. A task or file named on the item being written is found by its name; if the item is not known, ask the user which item it belongs to. When the user names a specific resource by its number or…

update_item_task

ChatGPT
Change the content of an existing task - its name, type, description or spec. Send only the fields that should change; anything omitted is left exactly as it is. Send an empty string to clear the description or spec. Use assign_item_task for the owner and update_item_task_time_left for the hours left (or to clear the estimate) - neither can be set here. Returns the task after the change.

update_item_task_time_left

ChatGPT
Update how many hours of work are left on a task. Project Simple recomputes the item's status from its tasks on every save, so dropping the last open task on a task type down to 0 can move the item's status forward on its own - the item's resulting status is returned alongside the task so there is no need to re-fetch it to find out. Omit hoursLeft to clear the task's estimate, leaving it unestimated. Returns the task and the item's status after the change.

update_sprint

ChatGPT
Change a sprint's dates, goals, utilization ratio or work-in-progress limit. Use it for "move the sprint end to Friday" or "set the utilization ratio to 70". Only the fields you pass change; everything else is left as it is. A sprint that has already started keeps its start date, and a sprint that has ended cannot be edited at all. New dates must not overlap another sprint of the same team. Returns the sprint as it now is.

update_time_off

ChatGPT
Change an existing time-off entry - who it covers, its dates, or its description. Send only the fields that should change; anything omitted is left exactly as it is, so there is no need to read the entry first and repeat the rest. Send an empty string to clear the description. Dates must be sent together - give both from and to, or neither. Same membership rule as scheduling: everyone the entry ends up covering has to be on a team that works on this project. Capacity of any sprint the old or new dates fall in is recalculated straight away. Returns the entry in full after the change.

update_todo_item

ChatGPT
Change a TODO item's name, description, owners, deadline, label or completion rule. Only the item's owners and its author may change it. Send only the fields to change; an empty string clears the description, label or deadline. ownerIds replaces the whole set of owners, and owners kept from before keep their completion state. Completion is not edited here - use resolve_todo_item and unresolve_todo_item; delete_todo_item drops the item. Returns the updated item in full.

view_file

ChatGPT
Look at what attached files hold - at most 4 per call: files on items, tasks, epics, goals, ideas, comments and TODO items, or files uploaded to the current AI Assistant conversation. Any type: an image comes back as a picture, scaled down to 1568 px on the longer side, whatever its size; text, code, CSV and logs as their text, from the beginning of a large file; Word, Excel, PowerPoint and the text files in a ZIP as their text, up to 50 MB; anything else - a PDF, audio, video - as the file itself, up to 20 MB, for a model that reads that type. Up to 95,000 characters of text per file. Call it only when the request needs what is in a file - to describe, summarize, compare or read it; never to attach, copy or move files, which needs only their ids. Pass fullResolution only to read small text a scaled image blurs; it sends images as they are, up to 20 MB. Returns a line per id saying whether it was shown, then the files.

vote_on_idea

ChatGPT
Cast, change or retract the signed-in user's vote on an idea. Pass SUPPORTED or OPPOSED to vote; omit voteType to retract the current user's vote. A user can be counted in only one vote bucket, so changing from SUPPORTED to OPPOSED removes the old vote automatically. Returns the updated idea.

Capabilities

Writes

Example Prompts

Click any prompt to copy it.

Similar Apps

Listed on MCP App Store

Listed on MCP App Store

Add this badge to your site or README. It links to this page.

HTML

<a href="https://mcp-app-store.keenethics-labs.com/app/project-simple"><img src="https://mcp-app-store.keenethics-labs.com/badges/listed-on-mcp-app-store.svg" alt="Listed on MCP App Store" width="220" height="44"></a>

Markdown

[![Listed on MCP App Store](https://mcp-app-store.keenethics-labs.com/badges/listed-on-mcp-app-store.svg)](https://mcp-app-store.keenethics-labs.com/app/project-simple)