routermcp
← Back to Blog

Research/CI MCP Pack (2025) — Perplexity, SearXNG, DeepWiki, Bright Data

• By RouterMCP Team

Run answer-first research, meta-search, deep wiki dives, and legally sourced web data collection via MCP.

Research workspace combining answer-first results with meta-search and wiki context.

Research/CI MCP Pack (2025) — Perplexity, SearXNG, DeepWiki, Bright Data

TL;DR: Triangulate answers across quality sources and keep an audit trail. Use Perplexity for answer-first, SearXNG for coverage, DeepWiki for long-tail docs, and Bright Data for programmatic collection where permitted.

Servers

  • Perplexity MCP (community). https://github.com/TimeSurgeLabs/mcp-server-perplexity
  • SearXNG MCP (community). https://github.com/tobychui/mcp-server-searxng
  • DeepWiki MCP (community). https://github.com/patarapolw/deepwiki
  • Bright Data MCP (community). https://github.com/Team-APJ/brightDataApi-mcp

Playbook

  1. Get an initial answer with citations (Perplexity).
  2. Expand coverage across engines (SearXNG) and collect canonical links.
  3. Deep dive a repo/topic (DeepWiki) and extract sections.
  4. Where allowed, queue Bright Data jobs; store results with provenance.

Internal links

  • Pack docs: /packs/research
  • Related posts: Security (01) for egress policies

FAQ Q: How do we avoid spam sources?
A: Maintain a domain allowlist and rank by authority (scholarly/government/primary docs first).

Schema

Checklist (fast)

  1. Intent. 2) Title/meta. 3) Slug. 4) TL;DR. 5) Playbook. 6) FAQ. 7) Links. 8) Images/alt. 9) Edit. 10) CTA.

CTA

  • Use the template: examples/packs/research.mcp.json and the “research notes” Notion template + citation exporter.