MCP App Store
by KeenEthics
Orchetta icon

Orchetta MCP Server for Claude and ChatGPT

By Orchetta

Overview

Orchetta is where a team manages its customer journeys: what breaks at each step, what is being done about it, who owns it, and whether the numbers moved. A journey is stages and steps; each step carries cards for the needs, pain points, opportunities, solutions, goals and metrics found there, and every card can point back to the research that raised it. Connected to Claude, that work is something you can ask about and act on in conversation. Ask: which journeys are in the workspace, where a given journey breaks, what is unmet, unaddressed or high priority, what each metric's latest reading is against its target, and who owns what. Decide: set a card's priority, status, scores and assignee, link a pain point to the solution that answers it, put the solutions worth showing the board in a list, and label work by the profile, product or brand it belongs to. Keep the map current: add a journey as a new one is found, add steps as the experience changes, and file the needs and pain points research turns up against the step they belong to. Documents and images in the file library are readable, so an interview write-up can be filed straight onto the map. Report: read a report, then rewrite, insert, move or remove its sections. Reports export to Word and PDF from the app. The connector acts as you, with the access your membership has, across every workspace you belong to. It calls no AI model of its own and spends none of your AI credits. Requires an Orchetta account; the 7-day trial needs no card.

How to connect

  1. Claude: Open the Orchetta listing in the Claude connectors directory and choose Connect. Claude listing Orchetta exposes 53 tools here.
  2. ChatGPT: Open the Orchetta listing in the ChatGPT app directory and choose Connect. ChatGPT listing Orchetta exposes 70 tools here.

Tools

add_atlas_stage

ChatGPT
Add one stage (column) to an existing atlas. Appends unless a position is given; bars already placed keep their columns.

add_evidence

ChatGPT
Back resources with evidence: what a customer said (quote), what someone saw or heard (observation), a web page (link), a file in the workspace's library (file), or a figure (data_point). Each item names the resources it backs, and one piece of evidence can back several — a quote behind a pain point and the need beneath it. Evidence already in the workspace saying the same thing from the same place is reused, so repeating a call adds nothing twice. A quote is the speaker's exact words, never tidied or paraphrased; a paraphrase is an observation. Attribute words to a role ("support lead") or a profile, never a person's name, and keep names, email addresses and other personal details out of the text. Only kinds marked evidence_enabled in list_resource_kinds take evidence. Up to 20 items per call.

add_flow_arrow

ChatGPT
Draw a flow arrow on a journey map from one action or touchpoint to another, so a sequence reads as a flow — across lanes too, a customer action handing off to a team's. Both ends must be records of a connectable kind (action, touchpoint; list_resource_kinds marks them) attached to steps of that journey. Drawing an arrow that's already there just sets its label.

add_metric_to_atlas_stage

ChatGPT
Pin a metric under a stage of an atlas, drawn as a chip showing the metric's latest reading. A placement only: the metric stays a record of the metric object (find one with list_resources kind "metric", or create one with create_resource) and keeps its readings wherever it's read.

add_records_to_list

ChatGPT
Put existing resources in a list. Records already listed stay put and count as added; ids that don't match anything are reported, not fatal.

add_resource_image

ChatGPT
Put an image on a resource by URL — a product screenshot on a solution, say. Only kinds with images enabled accept one; list_resource_kinds says which. The URL must be a public https:// address serving a PNG, JPEG, WebP or GIF.

add_stage

ChatGPT
Add one stage to an existing journey. Appends unless a position is given.

add_step

ChatGPT
Add one step to an existing stage. Appends unless a position is given.

apply_label

ChatGPT
Put a label — a profile, brand, product, or other labeling record — on a journey, stage, step, or resource. Labelling what's already labelled is a no-op. Labels never go on other labels.

ask_user

ChatGPT
Put a small questionnaire in front of the user: up to three questions, each with tappable choices and optionally a freeform "other" field. Use it when a decision forks the work — options beat prose the user has to retype. After calling it, END YOUR TURN with at most a short sentence; the answers arrive as the user's next message. Never call it when running unattended — nobody is there to answer.

attach_resource_to_step

ChatGPT
Attach an existing resource to a step on a journey map.

create_atlas

ChatGPT
Create an atlas: named stages that journeys are then plotted across. Check list_atlases first rather than duplicating one; place_journey_on_atlas fills it.

create_journey_map

ChatGPT
Create a journey with its stages and steps in one call. Prefer this over adding stages one at a time. For a conversational experience (a chatbot, a voice assistant, a support script) pass layout "conversation": stages are then the phases of the exchange and each step is one turn, with a speaker.

create_journey_view

ChatGPT
Save a view of a journey: a named way of reading the map that everyone in the workspace can pick, and that share links read through. A view changes what the map shows, never what's on it. Call list_journey_views first for the lane keys and for views that already exist.

create_list

ChatGPT
Create a list: a named, curated collection of existing resources — a shortlist of opportunities, the profiles to interview. Check list_lists first rather than duplicating one, then add_records_to_list fills it.

create_profile

ChatGPT
Create one profile: a persona this organization's experience serves. Call list_profiles first — a name already taken is refused rather than duplicated. A profile is a persona — one sort of person the experience serves, which the team then tags onto the journeys, steps, needs and opportunities that belong to them. A profile is a role with its own reasons, not a demographic slice and not a job title borrowed from an org chart. "First-time buyer" is a profile because they arrive knowing nothing and need proof before they'll commit; "25-34 year olds" is not, because nothing follows from it. The test for two profiles rather than one: would they take different paths through the same journey, or want different things at the same step? If they'd behave the same way, they are one profile with a broader description — merge them. Use 'customer' for someone who buys or uses what the organization offers, 'employee' for someone inside it who serves the experience, 'organization' when the persona really is a company or team rather than an individual, 'agent' for an AI agent acting on someone's behalf, and 'system' for software that takes part in the experience with nobody driving it.

create_resource

ChatGPT
Create a record of any object this organization tracks — a need, a metric, a brand, anything list_resource_kinds names — optionally attaching it to steps. Journeys and actions have their own tools, and labeling kinds (brand, product) go on the map with apply_label, never step_ids.

delete_atlas

ChatGPT
Delete an atlas: its stages and placements go with it, the journeys placed on it are untouched. Not undoable — check get_atlas first.

get_atlas

ChatGPT
One atlas in full: its stages in order (each a 1-based column) and every placed journey with the columns its bar spans.

get_document

ChatGPT
One report in full: title, subtitle, and its sections in order, each with the id the section tools take. A chart section also reports what its chart currently shows, so you can write about the numbers.

get_help_guide

ChatGPT
One help guide in full, as Markdown, by its slug from search_help_guides.

get_journey

ChatGPT
The full stage-and-step tree for one journey, including which resources are attached to each step.

get_journey_checks

ChatGPT
The workspace's checklist run against a journey: whether the map is complete enough to act on — described, staged, steps carrying actions and needs, needs scored and evidenced, a goal and a metric. Each check passes, fails or skips (nothing to judge), with a one-line reason. Omit journey_id for every journey's tally and failing checks, weakest first.

get_metric_readings

ChatGPT
A metric's readings, oldest first, between two days. list_resources shows only the latest 30; use this for the history behind them. Each reading says whether an integration synced it.

insert_document_section

ChatGPT
Add a section to a report, after a named section or at the end. A report holds at most 12 sections.

list_atlases

ChatGPT
List the organization's atlases — journey-of-journey charts that plot whole journeys as bars across shared stages, gantt-style.

list_change_sets

ChatGPT
Recent change sets, newest first: each is what one person's action, one MCP session, one assistant turn or one workflow run changed, which revert_change_set can undo together. Kept 90 days. Where undo isn't switched on for the workspace, answers with none and says so.

list_changes

ChatGPT
What changed in the workspace between two days, newest first: resources created, edited and deleted (each edited field with what it was and what it became, statuses included), evidence attached, drafts approved or rejected, and journeys and atlases created or changed. Give journey_id to keep to one journey and the resources on its steps. Defaults to the last 7 days.

list_documents

ChatGPT
The organization's reports — written documents edited in the app and exported to Word or PDF — newest first, with ids for get_document.

list_evidence

ChatGPT
The evidence behind one resource, or evidence across the workspace matching a search. Each piece says which resources it backs and where it came from.

list_files

ChatGPT
The workspace's file library — documents and images members uploaded for reference. Each entry says whether it has readable text; read one with read_file.

list_journey_views

ChatGPT
A journey's saved views — named ways of reading its map: which lanes show, in what order, and which cards pass the filters — and the lane keys the map has for a view to name.

list_journeys

ChatGPT
List the organization's journeys with their stage and step counts.

list_lists

ChatGPT
The organization's lists — curated collections of records — each with the record ids it already holds, so an add can skip what's already there.

list_members

ChatGPT
The organization's members. Call this to find the user id that assigns work — an assignee is a user, not a membership.

list_organizations

ChatGPT
List the organizations you belong to, with each organization's id (needed by every other tool), your role in it, and whether it admits MCP clients (mcp_enabled: false means every other tool will refuse it).

list_profiles

ChatGPT
List the organization's profiles (personas and segments).

list_resource_kinds

ChatGPT
The objects this organization tracks, with every attribute each one carries — keys, types, and the options its selects offer. Objects are per-organization, so call this before creating or filtering resources.

list_resources

ChatGPT
List the organization's resources of one kind, with their priority, status and scores.

move_document_section

ChatGPT
Move a section to sit after another, or to the top.

organization_overview

ChatGPT
Counts of everything in the organization plus its journeys. Call this first to orient.

place_journey_on_atlas

ChatGPT
Place a journey on an atlas as a bar spanning stages, or move a bar already there. A journey appears on an atlas at most once, so placing a placed journey moves its bar.

read_file

ChatGPT
Read one file from the library. Text and PDFs return their text in windows — pass the next_offset that comes back to continue a long file. Images are returned as the image itself, so you can look at screenshots and journey maps.

read_research

ChatGPT
One item from the research repository in full, by its ref from search_research: a highlight's exact words with who said them, in which interview and when; a note's text (a whole interview or transcript, a survey response's answers, the text of documents attached to it) followed by its fields, speakers and attachments; a doc or theme; or a project's overview with the refs of its notes and docs to read next. Long text comes in windows — pass the next_offset that comes back to read on. A workspace with no repository connected gets nothing and the reason.

remove_atlas_stage

ChatGPT
Take a stage (column) off an atlas. Nothing placed is lost: a bar that started on it moves to the first column, one that ended on it shrinks to its start. To merge two stages, remove one, rename the other, and re-place any bar that shrank.

remove_document_section

ChatGPT
Take one section out of a report.

remove_evidence

ChatGPT
Take a piece of evidence off one resource. It stays on any other resource it backs; evidence left backing nothing is deleted.

remove_flow_arrow

ChatGPT
Take a flow arrow off a journey map. The records at either end stay.

remove_journey_from_atlas

ChatGPT
Take a journey's bar off an atlas. The journey itself is untouched — an atlas holds placements, never the journeys themselves.

remove_label

ChatGPT
Take a label off a journey, stage, step, or resource.

remove_metric_from_atlas_stage

ChatGPT
Take a metric's chip off a stage of an atlas. The metric and its readings are untouched.

remove_metric_readings

ChatGPT
Delete a metric's hand-entered readings on the given days. A day an integration synced is left alone. Cannot be undone.

remove_stage

ChatGPT
Delete a stage and its steps. Records attached to those steps stay in the workspace and only come off this journey. Not undoable — get_journey first.

remove_step

ChatGPT
Delete a step. Records attached to it stay in the workspace and only come off this journey. Not undoable — get_journey first.

replace_document_section

ChatGPT
Rewrite one section of a report in place, keeping its position. Pass the whole section as it should read afterwards — kind, heading and content.

revert_change_set

ChatGPT
Undo a change set (list_change_sets, or the change_set_id a write returned). Without confirm it only reports what would be undone: tell the person, then call again with confirm true. Anything somebody changed since is left as it is. Where undo isn't switched on for the workspace, nothing changes and the answer says so.

search_help_guides

ChatGPT
Search Orchetta's own help guides — how the product works and how to do things in it. Answer how-do-I and where-is questions from these rather than from memory; an empty result returns the full table of contents instead.

search_research

ChatGPT
Search the workspace's connected research repository (Dovetail). Kinds: highlight (excerpts the team marked in interviews, matched by their words), note (Dovetail's "data": the interviews, transcripts, survey responses and notes themselves, highlighted or not, matched by what they say as well as their titles), doc (write-ups, by title), theme (patterns found across feedback) and project (the studies everything is kept in, by name). A study's name is a project, not words in its highlights: find the project first, then pass its id as project_id to search inside it; no query lists the newest. When next_cursor comes back, pass it as cursor for the next page. Each result's ref reads in full with read_research and backs resources through add_evidence's research_ref. A workspace with no repository connected gets no results and the reason.

show_page

ChatGPT
Bring a page of the app onto the user's screen, when they are following along — use it while guiding or touring, so they see what you describe. Changes nothing. Pages: home, journeys, atlas, portfolio, profiles, metrics, cycles, lists, boards. To show one specific journey, use get_journey instead — reading it brings its map up the same way.

update_atlas

ChatGPT
Rename or describe an atlas. Only the fields you pass are touched; stages and placements have their own tools.

update_atlas_stage

ChatGPT
Rename a stage of an atlas, or move it to another column. Bars anchored to the stage move with it.

update_document

ChatGPT
Change a report's title or subtitle. Only the fields you pass change.

update_evidence

ChatGPT
Change a piece of evidence. It changes on every resource it backs. Fields left out stay as they are; send an empty string to clear one.

update_journey

ChatGPT
Change an existing journey. Only the fields you pass are touched. Stages and steps have their own tools.

update_journey_view

ChatGPT
Change a saved view of a journey. Only the fields you pass are touched; lane lists you pass replace the view's.

update_resource

ChatGPT
Change an existing record's values. Only the keys you pass are touched, so setting a priority leaves a status alone. Journeys have their own tool.

update_stage

ChatGPT
Rename a stage of a journey, or move it to another position. Its steps and everything attached to them move with it.

update_step

ChatGPT
Change a step: its name, description or speaker, its position within its stage, or the stage it sits in (on the same journey). Only the fields you pass are touched; what is attached to the step stays attached.

whoami

ChatGPT
Confirm which Orchetta account this connection is authenticated as.

add_atlas_stage

Claude

add_metric_to_atlas_stage

Claude

add_records_to_list

Claude

add_resource_image

Claude

add_stage

Claude

add_step

Claude

apply_label

Claude

ask_user

Claude

attach_resource_to_step

Claude

create_atlas

Claude

create_journey_map

Claude

create_list

Claude

create_profile

Claude

create_resource

Claude

delete_atlas

Claude

get_atlas

Claude

get_document

Claude

get_help_guide

Claude

get_journey

Claude

insert_document_section

Claude

list_atlases

Claude

list_documents

Claude

list_files

Claude

list_journeys

Claude

list_lists

Claude

list_members

Claude

list_organizations

Claude

list_profiles

Claude

list_resource_kinds

Claude

list_resources

Claude

move_document_section

Claude

organization_overview

Claude

place_journey_on_atlas

Claude

read_file

Claude

remove_atlas_stage

Claude

remove_document_section

Claude

remove_journey_from_atlas

Claude

remove_label

Claude

remove_metric_from_atlas_stage

Claude

remove_stage

Claude

remove_step

Claude

replace_document_section

Claude

search_help_guides

Claude

show_page

Claude

update_atlas

Claude

update_atlas_stage

Claude

update_document

Claude

update_journey

Claude

update_resource

Claude

update_stage

Claude

update_step

Claude

whoami

Claude

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/orchetta"><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/orchetta)