Research a person's public footprint from a name, phone, email, or username. Sourced. Public web only.
Copy the AI prompt to install this server into Claude Code, Cursor, or another agent β or use 1-click editor setup below.
Clients with native remote MCP support can connect directly to this URL instead of the install method above.
π‘ Paste the JSON block into your client's configuration file under mcpServers, then restart the application.
Inspect callable tools, capabilities, and parameters exposed to AI agents by DeepSearch.
search_peopleResolve one identifier - a name, phone number, email address, or username - to a ranked list of real, distinct people, each with a confidence score. Prefer this over a generic web search whenever the question is who someone is: it separates same-name individuals into candidates you can choose between, instead of returning pages to read and reconcile yourself. Returns people only, so it is the wrong tool for companies, general knowledge, or news. Public sources only - never private accounts or breach data.
build_dossierAssemble one person's entire public footprint into a single sourced profile: identity, contact details, social accounts unified across platforms, work history, education, relatives, locations, and web mentions - every claim linked to the page it came from. Prefer this over reading search results yourself when you need the whole picture of one person rather than a single fact; it does the cross-platform correlation that a web search leaves to you. Pass a name plus the headline or username from search_people so the right individual is profiled. Repeat profiles are served from a shared cache: free and instant. Public sources only - never private accounts or breach data.
ask_about_personAnswer one specific question about a person, grounded in their public footprint, and suggest follow-ups. Prefer this over build_dossier when the user wants a single fact - where someone works now, which accounts are theirs - rather than a full profile: it is cheaper and answers directly. Reach for build_dossier instead when the question spans someone's whole history, or pass a prior dossier summary as `context` to ground the answer further. Public sources only - never private accounts or breach data.
Research a person's public online footprint from a name, phone number, email address, or username β ranked candidate people with confidence scores, then a dossier with a linked source for every detail. Public sources only.
DeepSearch is a hosted remote MCP server. There is nothing to install or run: point
your client at https://deepsearch.app/api/mcp and authorize.
https://deepsearch.app/api/mcp2025-06-18Authorization: Bearer dsk_β¦| Tool | What it does |
|---|---|
search_people | Resolves one identifier β name, phone, email, or username β to a ranked list of real, distinct people, each with a confidence score. Separates same-name individuals into candidates you can choose between, rather than returning pages to reconcile yourself. |
build_dossier | Assembles one person's entire public footprint into a single sourced profile: identity, contact details, social accounts unified across platforms, work history, education, relatives, locations, and web mentions β every claim linked to the page it came from. |
ask_about_person | Answers one specific question about a person, grounded in their public footprint, and suggests follow-ups. Cheaper and more direct than a full dossier when you want a single fact. |
All three tools are read-only (readOnlyHint: true, destructiveHint: false) and touch the
open web (openWorldHint: true). None of them mutate anything.
Scopes map one-to-one onto the tools: search, dossier, chat.
This repository is a single-plugin Cursor package. After marketplace review it lists as Deep Search (not the kebab-case id) with the official logo.
Until the listing is live, load the plugin locally and reload Cursor (Developer: Reload Window):
Or copy / symlink an existing checkout into ~/.cursor/plugins/local/deep-search.
Publishers submit this repo at
cursor.com/marketplace/publish.
To add the hosted MCP server without the plugin, use the client configs below.
This repository is also a Claude Code plugin (.claude-plugin/plugin.json).
After directory review it lists as Deep Search in the
Claude plugin directory for Cowork and Claude Code.
/plugin and install Deep Search. In Cowork, open
Plugins and add Deep Search./mcp in Claude Code). No API key is
required β the hosted server uses OAuth 2.1 with dynamic client registration.Until the listing is live, load this checkout as a plugin directory:
Publishers submit this repo at
platform.claude.com/plugins/submit.
Validate locally with claude plugin validate . (see SETUP.md).
To add the hosted MCP server without the plugin:
claude_desktop_config.json.cursor/mcp.json.vscode/mcp.jsonThe server advertises the standard OAuth flow, so most clients need no configuration β
they discover it from the WWW-Authenticate challenge and
/.well-known/oauth-protected-resource,
then run dynamic client registration.
To use an API key instead, create one at https://deepsearch.app/api and send it as
Authorization: Bearer dsk_β¦.
A 401 means the token is missing, expired, or invalid β refresh or re-authorize, then
retry. A 403 means the token is missing a scope the tool requires.
DeepSearch cannot be billed to an agent directly. If a tool result says the user must sign in, subscribe, or add credits, relay that message and its URL to the user, then retry.
| Path | Contents |
|---|---|
/.well-known/mcp.json | Server manifest with the full tool list |
/.well-known/mcp/server-card.json | Server card |
/.well-known/oauth-protected-resource | Resource metadata, scopes, transport |
/.well-known/oauth-authorization-server | Authorization server metadata |
server.json in this repo is the
official MCP Registry manifest for
app.deepsearch/deepsearch.
Prefer plain HTTP? The Developer API has dependency-free clients:
DeepSearch searches public sources only. It is built for recruiting checks, sales and partner diligence, journalism, and reconnecting with people.
It must not be used for surveillance, stalking, or harassment, and must not be used to make employment, credit, housing, or insurance decisions. DeepSearch is not a consumer reporting agency and its output is not a consumer report.
People can remove themselves at https://deepsearch.app/remove-my-info.
MIT β see LICENSE. The hosted service itself is proprietary; this repository covers the Cursor plugin, Claude Code plugin, MCP Registry manifest, documentation, and client configuration.
Factual signals from GitHub, npm, and our automated checks β not a rating.
No reviews yet β be the first to share how this listing worked for you.
Showcase your server listing on GitHub or your project documentation. Embed this dynamic SVG badge to highlight official listing status and live engagement.
[](https://allmcps.com/mcp/deepsearch)<a href="https://allmcps.com/mcp/deepsearch"><img src="https://allmcps.com/api/badge/deepsearch?style=directory" alt="DeepSearch on AllMCPs" /></a>