MCP Catalogs
Homegrainulator screenshot

grainulator

by grainulation·85·Score 47

Multi-pass research tool with claim-based investigation and confidence scoring for Claude Code.

ai-llmdeveloper-toolsproductivity
6
Forks
0
Open issues
1 mo ago
Last commit
2d ago
Indexed

Overview

Grainulator is a sophisticated research tool that performs multi-pass investigations, compiles typed claims, detects tensions between findings, and produces confidence-graded answers. It operates through a two-phase process: investigation (3 research passes) and compilation (7 analysis passes). The tool integrates with Claude Code as a plugin, providing both autonomous research capabilities and slash commands for controlled investigations.

Try asking AI

After installing, here are 5 things you can ask your AI assistant:

you:Comprehensive research with structured evidence gathering
you:Adversarial testing and validation of claims
you:Automated knowledge compilation and brief generation
you:What makes Grainulator different from regular chatbots?
you:How does the confidence scoring work?

When to choose this

Choose Grainulator for research-intensive projects requiring structured evidence gathering, conflict detection, and confidence scoring before making decisions.

When NOT to choose this

Don't choose Grainulator if you need rapid, simple answers or prefer a simpler research process without the complexity of multiple investigation passes and claim tension analysis.

Tools this server exposes

12 tools extracted from the README
  • /init

    Start a new research sprint

  • /research

    Multi-pass investigation with evidence gathering

  • /challenge

    Adversarial testing of a specific claim

  • /witness

    Corroborate a claim against an external source

  • /blind-spot

    Structural gap analysis

  • /brief

    Compiled decision brief

  • /present

    Presentation deck

  • /status

    Sprint dashboard

  • /pull

    Import knowledge from DeepWiki or Confluence

  • /sync

    Publish artifacts to Confluence

  • /calibrate

    Score predictions against actual outcomes

  • /resolve

    Adjudicate conflicts between claims

Comparable tools

semantic-search-mcpresearch-assistantknowledge-graph-mcpclaude-haystack

Installation

Installation

**Step 1** — Add the marketplace (one-time):

claude plugin marketplace add https://github.com/grainulation/grainulator.git

**Step 2** — Install:

claude plugin install grainulator

For Claude Desktop, add to claude_desktop_config.json:

{
  "mcpServers": {
    "grainulator": {
      "command": "npx",
      "args": ["-y", "@grainulation/grainulator", "serve-mcp"]
    }
  }
}

FAQ

What makes Grainulator different from regular chatbots?
Grainulator performs multi-pass research with typed claims and tension detection, producing confidence-graded answers rather than simple responses. It's a research sprint, not a conversation.
How does the confidence scoring work?
The confidence score (0-100) is calculated through 7 analysis passes evaluating evidence strength, type coverage, bias, gap analysis, and conflict detection.

On Hacker News

Recent discussion from the developer community.

Compare grainulator with

GitHub →

Last updated · Auto-generated from public README + GitHub signals.