research-agent-mcp

Multi-source research pipeline handling document analysis, stakeholder mapping, and evidence extraction. Part of the Adaptivearts.ai research-agent-mcps monorepo.

Tools

6 MCP tools composing the research pipeline

clarify

Decompose ambiguous queries into structured research questions with scoping constraints.

search

Multi-source document retrieval with relevance ranking and deduplication.

evaluate

Source credibility scoring, bias detection, and evidence strength assessment.

extract

Structured extraction of claims, entities, relationships, and supporting evidence.

synthesize

Cross-source synthesis with contradiction detection and confidence weighting.

report

Formatted research output with citations, evidence chains, and confidence levels.

Workflow

Sequential pipeline from ambiguity to structured output

Clarify Ambiguous input becomes structured research questions

Search Multi-source retrieval with relevance filtering

Evaluate Source credibility and evidence strength scoring

Extract Structured claims, entities, and relationships

Synthesize Cross-source merging with contradiction detection

Report Formatted output with citations and confidence levels

Integration Points

Works alongside sibling servers in the research-agent-mcps monorepo

research-notes-mcp

Structures raw notes from research output into themes and questions.

research-log-mcp

Logs citations and formats bibliographies from extracted sources.

evaluation-mcp (Evalla)

Rubric-based quality scoring of research outputs and evidence chains.

Additional Materials

Study materials coming soon.

Sibling

research-notes-mcp →

Sibling

research-log-mcp →

Prototype 01

Research Agent Suite →