MCP Configuration

Viben: Agent Swarm x Code Evolution

Viben is an AI-driven code iteration optimization and agent swarm orchestration platform. The MCP server is a key component of the platform, providing knowledge acquisition capabilities for Agent Swarm:

• FileEvo - Acquire best practices through academic papers and technical documentation to optimize code iteration
• Idea Generation - Gather inspiration from diverse data sources to spark innovative ideas
• Agent Swarm - Agent swarms can uniformly call MCP tools for knowledge search

---

Environment Variable Configuration

The Viben MCP server can be configured through environment variables. This page covers all available configuration options.

Environment Variables Overview

Variable	Description	Default
BROWSE_MCP_DOWNLOAD_PATH	Directory for saving downloaded content	./downloads
BROWSE_MCP_ENABLED_SOURCES	Enabled data sources (comma-separated)	All
BROWSE_MCP_DISABLED_SOURCES	Disabled data sources (comma-separated)	None

Download Path

Set the directory for saving downloaded content:

{
  "env": {
    "BROWSE_MCP_DOWNLOAD_PATH": "/path/to/downloads"
  }
}

If the directory does not exist, it will be created automatically.

Data Source Control

Data source control applies to all sources, including those provided by plugins.

Enable Only Specific Data Sources

Use BROWSE_MCP_ENABLED_SOURCES to create a whitelist:

{
  "env": {
    "BROWSE_MCP_ENABLED_SOURCES": "arxiv,pubmed,semantic,github"
  }
}

Only the listed data sources will be available. This works for both built-in and plugin data sources.

Disable Specific Data Sources

Use BROWSE_MCP_DISABLED_SOURCES to create a blacklist:

{
  "env": {
    "BROWSE_MCP_DISABLED_SOURCES": "ieee,scopus,springer,sciencedirect,twitter"
  }
}

All data sources except those listed will be available.

Priority Rules

• If BROWSE_MCP_ENABLED_SOURCES is set, it takes precedence
• If only BROWSE_MCP_DISABLED_SOURCES is set, all except listed sources are enabled
• If neither is set, all data sources are enabled by default

:::tip Recommended Configuration For most users, use BROWSE_MCP_DISABLED_SOURCES to disable paid data sources that require API keys:

{
  "env": {
    "BROWSE_MCP_DISABLED_SOURCES": "ieee,scopus,springer,sciencedirect,wos,acm,jstor"
  }
}

:::

API Keys

Academic Data Source API Keys

Variable	Service	How to Obtain
SEMANTIC_SCHOLAR_API_KEY	Semantic Scholar	Get API Key
CORE_API_KEY	CORE	Get API Key
IEEE_API_KEY	IEEE Xplore	Get API Key
SCOPUS_API_KEY	Scopus	Get API Key
SPRINGER_API_KEY	Springer Link	Get API Key
SCIENCEDIRECT_API_KEY	ScienceDirect	Get API Key
WOS_API_KEY	Web of Science	Requires institutional subscription

Plugin API Keys (Social Media)

Variable	Service	How to Obtain
GITHUB_TOKEN	GitHub	Personal Access Token
TWITTER_BEARER_TOKEN	Twitter/X	Twitter Developer Platform
ZHIHU_API_KEY	Zhihu	Contact platform
XIAOHONGSHU_API_KEY	Xiaohongshu	Contact platform

Free vs Paid Data Sources

Free Data Sources (no API key required):

• arxiv, pubmed, pmc, biorxiv, medrxiv
• google_scholar, iacr, crossref, researchgate
• github (with rate limits)

Free but Optional API Key (higher limits with key):

• semantic - Works without key, higher limits with SEMANTIC_SCHOLAR_API_KEY
• core - Requires CORE_API_KEY
• github - 60 requests/hour without key, 5000 requests/hour with key

Paid Data Sources (require API key):

• ieee - Requires IEEE_API_KEY
• scopus - Requires SCOPUS_API_KEY
• springer - Requires SPRINGER_API_KEY
• sciencedirect - Requires SCIENCEDIRECT_API_KEY
• wos - Requires WOS_API_KEY and institutional subscription
• twitter - Requires TWITTER_BEARER_TOKEN

Available Data Source Names

Built-in Academic Data Sources

Data Source Name	Type	Description
arxiv	Free	Physics, math, computer science preprints
pubmed	Free	MEDLINE biomedical literature
pmc	Free	PubMed Central full-text repository
biorxiv	Free	Biology preprint server
medrxiv	Free	Health sciences preprint server
google_scholar	Free	Google Scholar search
iacr	Free	IACR cryptography preprints
semantic	Free	Semantic Scholar (optional API key)
crossref	Free	CrossRef DOI metadata
core	Free	CORE open access (requires API key)
ieee	Paid	IEEE Xplore digital library
scopus	Paid	Elsevier Scopus database
springer	Paid	Springer publications
sciencedirect	Paid	Elsevier ScienceDirect
wos	Paid	Web of Science
acm	Paid	ACM Digital Library
jstor	Paid	JSTOR archive
researchgate	Free	ResearchGate social network

Plugin Data Sources (Social Media)

These data sources require installing browse-mcp-plugin-social-media:

Data Source Name	Type	Description
github	Free	GitHub repositories and code
twitter	Paid	Twitter/X posts (requires API key)
zhihu	Free	Zhihu Q&A articles
xiaohongshu	Free	Xiaohongshu posts

Configuration Examples

Minimal Configuration (Free Sources Only)

{
  "mcpServers": {
    "browse-mcp": {
      "command": "python",
      "args": ["-m", "browse_mcp"],
      "env": {
        "BROWSE_MCP_DOWNLOAD_PATH": "./downloads"
      }
    }
  }
}

With Semantic Scholar API Key

{
  "mcpServers": {
    "browse-mcp": {
      "command": "python",
      "args": ["-m", "browse_mcp"],
      "env": {
        "SEMANTIC_SCHOLAR_API_KEY": "your-api-key-here",
        "BROWSE_MCP_DOWNLOAD_PATH": "./downloads"
      }
    }
  }
}

Academic + Social Media

{
  "mcpServers": {
    "browse-mcp": {
      "command": "python",
      "args": ["-m", "browse_mcp"],
      "env": {
        "SEMANTIC_SCHOLAR_API_KEY": "your-key",
        "GITHUB_TOKEN": "ghp_your_github_token",
        "BROWSE_MCP_DOWNLOAD_PATH": "./downloads"
      }
    }
  }
}

Research Focus (Medical/Biology)

{
  "mcpServers": {
    "browse-mcp": {
      "command": "python",
      "args": ["-m", "browse_mcp"],
      "env": {
        "BROWSE_MCP_ENABLED_SOURCES": "pubmed,pmc,biorxiv,medrxiv,semantic",
        "SEMANTIC_SCHOLAR_API_KEY": "your-key",
        "BROWSE_MCP_DOWNLOAD_PATH": "./medical-papers"
      }
    }
  }
}

Computer Science Focus

{
  "mcpServers": {
    "browse-mcp": {
      "command": "python",
      "args": ["-m", "browse_mcp"],
      "env": {
        "BROWSE_MCP_ENABLED_SOURCES": "arxiv,semantic,ieee,acm,github",
        "SEMANTIC_SCHOLAR_API_KEY": "your-key",
        "IEEE_API_KEY": "your-key",
        "GITHUB_TOKEN": "ghp_your_token",
        "BROWSE_MCP_DOWNLOAD_PATH": "./cs-papers"
      }
    }
  }
}

Social Media Only

{
  "mcpServers": {
    "browse-mcp": {
      "command": "python",
      "args": ["-m", "browse_mcp"],
      "env": {
        "BROWSE_MCP_ENABLED_SOURCES": "github,twitter,zhihu",
        "GITHUB_TOKEN": "ghp_your_token",
        "TWITTER_BEARER_TOKEN": "your_bearer_token",
        "BROWSE_MCP_DOWNLOAD_PATH": "./social-content"
      }
    }
  }
}

Troubleshooting

Data Source Unavailable

If a data source is unavailable:

1. Check if it's in BROWSE_MCP_DISABLED_SOURCES
2. Check if BROWSE_MCP_ENABLED_SOURCES is set and includes the data source
3. For paid data sources, verify the API key is set
4. For plugin data sources, verify the plugin is installed: pip show browse-mcp-plugin-social-media

API Rate Limits

If you encounter rate limits:

• Semantic Scholar: Add SEMANTIC_SCHOLAR_API_KEY for higher limits
• CORE: Get a free API key from core.ac.uk
• Google Scholar: May be rate limited; use other data sources as alternatives
• GitHub: Add GITHUB_TOKEN to increase from 60 requests/hour to 5000 requests/hour

Downloaded Files Missing

If downloaded files are missing:

1. Check if BROWSE_MCP_DOWNLOAD_PATH is writable
2. Verify the directory exists or can be created
3. Check error messages in the response

Plugin Not Loading

If plugin data sources are unavailable:

1. Verify the plugin is installed: pip list | grep browse-mcp
2. Check for loading errors: browse-mcp --debug
3. Verify entry points: python -c "from stevedore import ExtensionManager; print([e.name for e in ExtensionManager('browse_mcp.searchers')])"

Next Steps

• Client Configuration - Configure MCP client
• browse_search Tool - Learn search parameters
• Plugin Configuration - Advanced plugin settings