search_workspaceLABSCO SUMMARY
Your research corpus, searchable and writable from the chat.
The reach is the point: search the workspace, open the transcript behind a quote, then create the doc that cites it, without pasting anything between windows. The assistant inherits your permissions exactly — it cannot open a project you cannot — and anything it creates is recorded as yours.
What it is
Dovetail's server connects an AI assistant to your Dovetail workspace. It searches across projects, docs, highlights and channels, reads content as markdown, and creates new projects, docs, data and highlights on your behalf.
What you get
- Workspace-wide search, plus projects, folders and folder contents listed and read
- Data entries and docs read as markdown, and new ones created
- Highlights retrieved, and a new one created on an audio or video transcript
- Channels: their raw data points, their themes, and new data sent in
- Workspace members, contacts, tags, custom fields, and file metadata with short-lived download links
- Comments on a doc listed, read and posted
- A hosted endpoint over Streamable HTTP, or a local server over stdio for clients that only support stdio
- Built-in integrations in ChatGPT, Claude, Figma and Copilot, with nothing to configure
Requirements
A Dovetail workspace. The hosted endpoint authenticates through your browser over OAuth, or with a Dovetail API token as a bearer header — your client needs to support one of the two. The local server takes an API token.
Setup effort
Follow the published steps — Prerequisites: Node.js 22 or higher, and a Dovetail API token
USE CASES
01
Answering a question about what customers actually said
Search across the workspace, then open the specific transcripts behind the results.
02
Finding evidence for a hypothesis
Pull highlights from the relevant project and read the docs already published around them.
03
Writing a finding up where the evidence lives
Create the doc in the right project and mark the transcript passage it rests on.
04
Bringing outside feedback in
Send a new data point to the channel that tracks it, and read the themes it lands under.
05
Starting a project from a template
List the templates, create the project, and file it in the folder where it belongs.
40 tools
Listed in project docs. Your client may show a different list.
get_dovetail_projectsget_projectcreate_projectlist_project_templateslist_foldersget_folderget_folder_contentscreate_folderlist_project_dataget_project_dataget_data_contentcreate_dataget_project_highlightsget_highlightcreate_transcript_highlightlist_docsget_docget_doc_contentcreate_doclist_doc_commentsget_doc_commentcreate_commentlist_channelsget_channellist_channel_dataget_channel_datumlist_channel_themescreate_channel_datumlist_usersget_userlist_contactsget_contactlist_tagsget_tagcreate_taglist_fieldsget_fieldget_filedownload_fileBEFORE YOU INSTALL
What to know first
Tools you get
40 tools named in the project's documentation.
Requires
- Node.js 22 or later
- DOVETAIL_API_TOKEN in your environment
Where it runs
Of the 4 documented clients, 1 runs the server on your machine and 3 connect to a hosted endpoint.
INSTALLATION
Connect your client
Server setup
Setup
Prerequisites: Node.js 22 or higher, and a Dovetail API token
Option 1 — download the latest index.js from https://github.com/dovetail/dovetail-mcp/releases/latest/download/index.js, then go straight to configuration
Option 2 step 1 — git clone https://github.com/dovetail/dovetail-mcp.git && cd dovetail-mcp
Option 2 step 2 — yarn install
Option 2 step 3 — yarn build (must run before the next step; dist/index.js does not exist until the build completes)
Option 2 step 4 — realpath dist/index.js | pbcopy to capture the path for the config
Add the mcpServers block to Claude / Cursor (Cursor: Settings → Tools & Integrations → Add Custom MCP), pasting the index.js path and your token
Hosted endpoint
ENDPOINT URL
https://dovetail.com/api/mcpAdd this hosted endpoint to your MCP client using the credentials listed under Requirements.
MCP Clients
Claude Desktop
CLAUDE DESKTOP · JSON
{
"mcpServers": {
"dovetail-mcp": {
"command": "node",
"args": [
"<YOUR_INDEX_JS_FILE>"
],
"env": {
"DOVETAIL_API_TOKEN": "<YOUR_TOKEN>"
}
}
}
}Cursor
CURSOR · JSON
{
"mcpServers": {
"dovetail-mcp": {
"url": "https://dovetail.com/api/mcp",
"headers": {
"Authorization": "Bearer <YOUR_TOKEN>"
}
}
}
}Windsurf
WINDSURF · JSON
{
"mcpServers": {
"dovetail-mcp": {
"serverUrl": "https://dovetail.com/api/mcp",
"headers": {
"Authorization": "Bearer <YOUR_TOKEN>"
}
}
}
}Microsoft Copilot Studio
In Copilot Studio, open the MCP onboarding wizard
Server URL:
MICROSOFT COPILOT STUDIO · URL
https://dovetail.com/api/mcpAuthentication: API key — Type: Header, Header name: Authorization
API key value: Bearer <YOUR_TOKEN>, using your Dovetail API token
