MCP Setup

Quick Config

Add c4reqber to your MCP client configuration:

copy{
  "mcpServers": {
    "c4reqber": {
      "command": "c4reqber",
      "args": ["serve", "--mcp"],
      "env": {
        "OPENROUTER_API_KEY": "${OPENROUTER_API_KEY}",
        "ARXIV_EMAIL": "your-email@example.com"
      }
    }
  }
}

Claude Desktop

copy{
  "mcpServers": {
    "c4reqber": {
      "command": "blast",
      "args": ["serve", "--mcp"],
      "env": {
        "OPENROUTER_API_KEY": "${OPENROUTER_API_KEY}"
      }
    }
  }
}

Available Tools

c4reqber exposes 21 MCP tools for cognitive operations (blast_* + c4_* families).

c4_fingerprint

Classify a problem into the C4 cognitive space (Z₃³, 27 states).

• Input: problem_text (string)
• Output: {state: (t, s, a), confidence: float, reasoning: string}

triz_contradiction

Solve engineering contradictions using the TRIZ 39×39 matrix.

• Input: improving_param (int 1-39), worsening_param (int 1-39)
• Output: {principles: [int], descriptions: [string], examples: [string]}

pattern_simulate

Run a scientific simulation pattern against a hypothesis.

• Input: pattern (string), params (dict), hypothesis (string)
• Output: {metrics: dict, charts: [bytes], execution_time_ms: float}
• Engines: Newton, TorchSim, JaxSim, Schr, vast.ai

causal_analyze

Discover causal relationships using Pearl's do-calculus.

• Input: variables (list), observations (list), query (string)
• Output: {graph: dict, do_effect: float, counterfactuals: [dict]}

paradigm_detect

Detect paradigm shifts in scientific literature.

• Input: domain (string), time_range (years), threshold (float)
• Output: {anomalies: [dict], contradictions: [dict], shift_probability: float}

lean4_verify

Formally verify a discovery using Lean 4 theorem prover.

• Input: hypothesis (string), premises (list)
• Output: {lean_file: string, verification_status: string, proof_log: string}
• Requires: Lean 4 toolchain installed locally

knowledge_search

Unified search across 51 federated knowledge sources.

• Input: query (string), sources (list), max_results (int)
• Output: {results: [{source, title, url, abstract, license}], total: int}

bayesian_update

Update beliefs using Bayesian inference.

• Input: prior (dict), likelihood (dict), method (MCMC/BMA)
• Output: {posterior: dict, model_probabilities: dict, diagnostics: dict}

system_dynamics

Simulate feedback loops using Stock-Flow DSL.

• Input: stocks (list), flows (list), params (dict), time_range (years)
• Output: {trajectory: [float], archetype: string, leverage_points: [dict]}

abductive_infer

Generate and rank explanatory hypotheses (Inference to the Best Explanation).

• Input: observations (list), domain_context (string)
• Output: {hypotheses: [{text, score, explanation}], ranking: [int]}

Registry Listings

c4reqber MCP server will be listed on major registries:

Registry	Status
modelcontextprotocol.io	Coming soon
Smithery.ai	Coming soon
PyPI	Coming soon