Upriver MCP
Provides AI applications with real-time, evidence-backed context on creators, audiences, brands, trends, and sponsorships, including breakout topic search and browsing tools.
README
<p align="center"> <img src="https://mcp.upriver.ai/icon.png" width="96" height="96" alt="Upriver" /> </p>
<h1 align="center">Upriver MCP</h1>
<p align="center"><em>Real-time context on creators, audiences, brands, trends, and sponsorships — for AI applications.</em></p>
About Upriver
Upriver provides an API that turns what's happening online into real-time, reliable context for AI applications. It delivers evidence-backed signals — each grounded in real sources — across:
- Creators — relevant creators and profiles
- Audiences — personas, interests, language, and humor
- Brands — brands and how they're being discussed
- Products — product mentions and use cases
- Trends — topics and formats gaining attention
- Sponsorships — brand–creator relationships and placements
This repository is the home of Upriver's Model Context Protocol (MCP) server, which brings that context to any MCP-compatible AI assistant or agent.
Connect
| Endpoint | https://mcp.upriver.ai/mcp |
| Transport | Streamable HTTP |
| Auth | Sign in with Upriver (OAuth), or send your Upriver API key in the X-API-Key header |
It's a remote, hosted server — nothing to install or run. OAuth sign-in works in
every client. The X-API-Key header is supported by Claude Code, Cursor, VS Code,
Cline, Windsurf, Codex CLI, Goose, and Zed — but not by Claude's or ChatGPT's
connector UIs, which are OAuth-only.
Claude Code (CLI)
claude mcp add --transport http upriver https://mcp.upriver.ai/mcp
To use an API key instead of OAuth, pass it as a header:
claude mcp add --transport http upriver https://mcp.upriver.ai/mcp \
--header "X-API-Key: YOUR_UPRIVER_API_KEY"
Cursor
Add to ~/.cursor/mcp.json (global) or .cursor/mcp.json (per-project), then connect
it in Settings → Tools & Integrations to sign in:
{ "mcpServers": { "upriver": { "url": "https://mcp.upriver.ai/mcp" } } }
For API-key auth, add "headers": { "X-API-Key": "YOUR_UPRIVER_API_KEY" }.
VS Code (GitHub Copilot)
VS Code uses a different shape — top-level servers and a required "type": "http".
Add to .vscode/mcp.json (or run MCP: Add Server):
{ "servers": { "upriver": { "type": "http", "url": "https://mcp.upriver.ai/mcp" } } }
For API-key auth, add "headers": { "X-API-Key": "YOUR_UPRIVER_API_KEY" }.
Claude Desktop & claude.ai
In Claude, go to Customize → Connectors → "+" → Add custom connector, name it
Upriver, paste https://mcp.upriver.ai/mcp, then Connect and sign in.
(Claude's connector is OAuth-only — there's no field for an API key, and don't add a
raw url entry to claude_desktop_config.json; the desktop app rejects it.)
Cline
Use the Remote Servers tab (Transport: Streamable HTTP), or edit
cline_mcp_settings.json — the type field is required, or Cline falls back to legacy
SSE and fails:
{ "mcpServers": { "upriver": { "type": "streamableHttp", "url": "https://mcp.upriver.ai/mcp" } } }
Windsurf
Edit ~/.codeium/windsurf/mcp_config.json — Windsurf's field is serverUrl:
{ "mcpServers": { "upriver": { "serverUrl": "https://mcp.upriver.ai/mcp" } } }
<details> <summary><b>More clients</b> — Codex CLI, ChatGPT, Goose, Zed</summary>
Codex CLI (OpenAI)
codex mcp add upriver --url https://mcp.upriver.ai/mcp --header X-API-Key=YOUR_UPRIVER_API_KEY
ChatGPT — Enable Developer mode (Settings → Apps & Connectors → Advanced), then
Settings → Connectors → Create, name it Upriver, paste https://mcp.upriver.ai/mcp,
choose OAuth, and sign in. (No custom-header field — OAuth only in the UI.)
Goose — add to ~/.config/goose/config.yaml:
extensions:
upriver:
type: streamable_http
uri: https://mcp.upriver.ai/mcp
enabled: true
headers:
X-API-Key: "${UPRIVER_API_KEY}"
Zed — add to settings.json under context_servers:
{ "context_servers": { "upriver": { "url": "https://mcp.upriver.ai/mcp", "headers": { "X-API-Key": "YOUR_UPRIVER_API_KEY" } } } }
</details>
Any other MCP client
Most clients use the Cursor shape above (mcpServers + url, optional headers).
For agents that read install manifests, see llms-install.md.
JetBrains IDEs: MCP works, but the IDE currently drops custom request headers and doesn't yet support OAuth for remote servers, so Upriver's
X-API-Keyauth can't connect there yet — we'll list it once JetBrains ships the fix.
Tools available today
The MCP server currently exposes Upriver's breakout topics — emerging subjects gaining traction online across tech, sports, and politics, each returned with grounded source citations so the model reasons from real evidence instead of guessing.
Both tools are read-only. Each returns up to limit topics (default 10,
max 25), every topic carrying up to 5 source citations, plus a has_more flag
when more topics exist beyond the page.
breakout_search_topics
Search breakout topics by keyword.
| Parameter | Type | Description |
|---|---|---|
query |
string · required | Natural-language search, e.g. "AI video effects". |
vertical |
enum | Restrict to tech, sports, or politics. |
limit |
integer | 1–25 (default 10). |
citation_sources |
array | Restrict citations to any of news, reddit, twitter. |
breakout_list_topics
Browse breakout topics gaining traction right now.
| Parameter | Type | Description |
|---|---|---|
vertical |
enum | Restrict to tech, sports, or politics. |
status |
enum | Lifecycle filter: active (default), emerging, trending, declining, detected, or all. |
limit |
integer | 1–25 (default 10). |
citation_sources |
array | Restrict citations to any of news, reddit, twitter. |
Each topic includes its name, vertical, lifecycle status, a momentum signal, and its top citations (title, URL, source category, and publish time).
More on the way
Breakout topics are the first slice of the Upriver API to land as MCP tools. The full API — Creators, Audiences, Brands, Products, Trends, and Sponsorships — is documented at docs.upriver.ai, and we're rolling those endpoints out as MCP tools. Watch this repository for new tools as they ship.
Access
Upriver MCP access is provisioned per account. Request access at dashboard.upriver.ai or email support@upriver.ai. Once your account is enabled, the same Upriver sign-in (or API key) works across every MCP client.
Links
- Docs — https://docs.upriver.ai
- Dashboard — https://dashboard.upriver.ai
- Website — https://upriver.ai
License
The contents of this repository — connection documentation and listing metadata — are released under the MIT License.
The hosted Upriver MCP service and the Upriver API are operated by Potato Laboratories, Inc. d/b/a Upriver and are governed by Upriver's Terms of Service. The MIT License covers this repository only — not the service, nor the data it returns. "Upriver" and the Upriver logo are trademarks of Potato Laboratories, Inc. and are not covered by the MIT License.
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.