UniFuncs MCP Server

by UniFuncs

10 stars
417 downloads
Not rated
GitHub

About

UniFuncs MCP Server provides access to the UniFuncs API through the Model Context Protocol (MCP), offering web search, web reader, deep search (sync and async), and deep research capabilities. It is designed for developers building AI applications that need real-time or in-depth…

Details

Author
UniFuncs
GitHub stars
10
Downloads
417
Categories
Other

- Real-time web search with freshness and pagination.
- Clean content extraction from web pages.
- Synchronous and asynchronous deep search.
- Comprehensive deep research with custom parameters.
- Multiple output formats: JSON, Markdown, Text.

Setting up with Highlight

This MCP is not yet compatible with Highlight’s one-click setup. However, you can still use it with Highlight by following these steps:

  1. Download and install Highlight from highlightai.com/download
  2. Navigate to the plugins tab and select "Add Custom Plugin"
  3. Configure the plugin with the settings below
    Plugin Name UniFuncs MCP Server
    Command (node, npx, python, etc.)

    Please refer to the README for specific instructions on how to obtain API keys or other required environment variables.

  4. Enable "Start Automatically" if you want the plugin to start when Highlight launches

From the repository

Install and configure via NPX (STDIO) using the @unifuncs/ufn-mcp-server package, setting the UNIFUNCS_API_KEY environment variable. Alternatively, build from source or run as an SSE server with the --sse flag or UNIFUNCS_SSE_SERVER=true. Tools are invoked by sending JSON-formatted requests to the MCP client.

Claude Desktop / Cursor

Paste into your MCP client config file to install this server.

{
    "mcpServers": {
        "unifuncs mcp server": {
            "unifuncs": {
                "command": "npx",
                "args": [
                    "-y",
                    "@unifuncs/ufn-mcp-server"
                ],
                "env": {
                    "UNIFUNCS_API_KEY": "<YOUR UNIFUNCS API KEY>"
                }
            }
        }
    }
}

McpServers

{
    "unifuncs": {
        "command": "npx",
        "args": [
            "-y",
            "@unifuncs/ufn-mcp-server"
        ],
        "env": {
            "UNIFUNCS_API_KEY": "<YOUR UNIFUNCS API KEY>"
        }
    }
}

UniFuncs MCP Server

MCP Server for the UniFuncs API - Enhanced with Deep Search and Deep Research capabilities

Features

This MCP server provides access to the following UniFuncs APIs:

1. Web Search (web-search)

Real-time web search with comprehensive results - Search across the internet with keywords - Filter by freshness (Day/Week/Month/Year) - Pagination support (1-50 results per page) - Multiple output formats (JSON/Markdown/Text)

2. Web Reader (web-reader)

Extract detailed content from web pages - Clean content extraction - Optional image inclusion - Link summary support - Markdown output

3. Deep Search - Sync (deep-search-sync)

Real-time deep search with immediate results - Model: S3 / S3 Pro - Streaming support - Instant response

4. Deep Search - Async (deep-search-create-task + deep-search-query-task)

Asynchronous deep search for complex queries - Create task and get task_id immediately - Poll for task status and results - Background processing for large queries

5. Deep Research (deep-research-create-task + deep-research-query-task)

Comprehensive deep research capabilities - Models: U3, U3 Pro - Customizable research parameters: - Introduction: Set researcher persona - Reference style: link/number/footnote - Max depth: Up to 50 iterations (25 recommended) - Domain scope: Limit search to specific domains - Domain blacklist: Exclude specific domains - Custom output prompts - Important URLs, keywords, and prompts - Async task management

Setup

API Key

Get a UniFuncs API Key:
https://unifuncs.com/account

NPX (STDIO)

{
    "mcpServers": {
        "unifuncs": {
            "command": "npx",
            "args": [
                "-y",
                "@unifuncs/ufn-mcp-server"
            ],
            "env": {
                "UNIFUNCS_API_KEY": "sk-"
            }
        }
    }
}

Note: the npx experience reflects the latest package published to npm, not necessarily the current main branch. If the published npm version is behind main, use the source build instructions below to reproduce the latest merged toolset.

Source Build (Local)

git clone https://github.com/UniFuncs/ufn-mcp-server.git
cd ufn-mcp-server
yarn install
npm run build
UNIFUNCS_API_KEY=sk- node build/index.js

Example Codex MCP config:

[mcp_servers.unifuncs-local]
command = "node"
args = ["/absolute/path/to/ufn-mcp-server/build/index.js"]

[mcp_servers.unifuncs-local.env]
UNIFUNCS_API_KEY = "sk-"

SSE Server

For SSE transport, set the environment variable:

export UNIFUNCS_SSE_SERVER=true
export UNIFUNCS_SSE_SERVER_PORT=5656  # Optional, default is 5656

Or use the --sse flag:

npx @unifuncs/ufn-mcp-server --sse

Tool Reference

web-search

Query: Search keywords
Freshness: Day | Week | Month | Year (optional)
Page: Page number, default 1 (optional)
Count: Results per page, 1-50, default 10 (optional)
Format: json | markdown | text, default json (optional)

web-reader

URL: Page URL to read
Format: markdown (optional)
IncludeImages: boolean (optional)
LinkSummary: boolean (optional)

deep-search-sync

Model: s3 / s3-pro (default: s3)
Messages: Array of {role: "user"|"assistant"|"system", content: string}
Stream: boolean (default: false)

deep-search-create-task

Model: s3 / s3-pro (default: s3)
Messages: Array of {role: "user"|"assistant"|"system", content: string}
Returns: task_id for querying status

deep-search-query-task

Task_ID: Task ID from create_task
Returns: Task status, progress, and results when completed

deep-research-create-task

Model: u3 | u3-pro (default: u3)
Content: Research question/topic
Introduction: Researcher persona (optional)
Reference_Style: link | number | footnote (default: link)
Generate_Summary: boolean (default: false)
Max_Depth: 1-50 (default: 25, recommended)
Domain_Scope: Comma-separated domains (optional)
Domain_Blacklist: Comma-separated domains to exclude (optional)
Output_Prompt: Custom output template (optional)
Important_URLs: Comma-separated URLs (optional)
Important_Keywords: Comma-separated keywords (optional)
Important_Prompt: Important prompt content (optional)
Push_To_Share: boolean (default: false)
Set_Public: boolean (default: false)
Returns: task_id for querying status

deep-research-query-task

Task_ID: Task ID from create_task
Returns: Task status, progress, and results when completed

Examples

Web Search

{
  "query": "OpenClaw AI",
  "count": 5,
  "format": "json"
}

Deep Search Async

// Create task
{
  "model": "s3",
  "messages": [
    { "role": "user", "content": "What are the latest developments in AI?" }
  ]
}

// Query task (use returned task_id)
{
"task_id": "3aff2a91-7795-4b73-8dab-0593551a27a1"
}

Deep Research

// Create research task
{
  "model": "u3",
  "content": "Analyze the impact of AI on healthcare",
  "max_depth": 25,
  "domain_scope": "arxiv.org, nature.com",
  "generate_summary": true
}

// Query research task
{
"task_id": "research-task-id-here"
}

Pricing

- Web Search: Pay per request
- Web Reader: Pay per request
- Deep Search: Pay per token usage
- Deep Research:
- U3: M Tokens
- U3 Pro: M Tokens

For detailed pricing, visit: https://unifuncs.com/pricing

Support

- Documentation: https://unifuncs.com/api
- GitHub: https://github.com/UniFuncs/ufn-mcp-server
- Email: service@unifuncs.com
- WeChat: unifuncs

Maintainer Release Checklist

Before documenting new tools in the npx setup section, make sure the npm package has been published and smoke-tested.

1. Bump the package version.
2. Install dependencies and run npm run build.
3. Publish the package to npm.
4. Verify npm view @unifuncs/ufn-mcp-server version matches the intended release.
5. Verify npx -y @unifuncs/ufn-mcp-server exposes the same tools described in this README.
6. Update the changelog and README after the npm release is available.

License

MIT

Changelog

v0.1.0 (2026-02-25)

- Added Deep Search sync API support - Added Deep Search async API (create_task + query_task) - Added Deep Research async API (create_task + query_task) - Enhanced documentation with all tool references - Improved error handling

v0.0.6

- Initial release with web-search and web-reader
No reviews yet — be the first

Sign in to leave a review

Use Google, GitHub, or an email account so ratings stay tied to real people.

Email sign in

No reviews posted yet.