Reference

API Reference

Blast commands, REST endpoints, Python SDK, and 21 MCP tools. Everything you need to integrate.

Blast Commands

The fastest way to use c4reqber from the command line. All commands support --help for detailed options.

$ blast solve "your research question"
$ blast turbo "complex multi-step problem"
$ blast flash "quick check"

Command	Description	Latency
`blast solve`	Standard discovery pipeline — full Z₃³ traversal	~5s
`blast turbo`	Accelerated path — skips low-relevance states	~2s
`blast flash`	Instant answer — single best engine	<1s
`blast turbofactory`	Batch processing from file	~1s/query

REST API Endpoints (v5.6.0)

Base URL: /api/v1

GET /health Service health check

POST /solve Submit a problem for discovery

GET /solve/{job_id} Get discovery results

POST /agenda Create a research agenda

GET /agenda/{agenda_id} Get agenda status and results

POST /verify Verify a hypothesis against knowledge sources

GET /engines List available simulation engines

GET /sources List configured knowledge sources

Example: Solve a Problem

curl -X POST https://api.c4reqber.org/v1/solve \\
  -H "Authorization: Bearer $TOKEN" \\
  -H "Content-Type: application/json" \\
  -d '{
    "query": "Find novel approaches to protein folding",
    "engines": ["biology", "optimization"],
    "max_depth": 3
  }'

Example: Create Agenda

curl -X POST https://api.c4reqber.org/v1/agenda \\
  -H "Authorization: Bearer $TOKEN" \\
  -d '{
    "topic": "Quantum computing applications in drug discovery",
    "duration_days": 30,
    "milestones": 5
  }'

Python SDK

from c4reqber import Client

client = Client(api_key="your-key")

# Solve a problem
result = client.solve("Novel battery chemistry")
print(result.discoveries)

# Create a research agenda
agenda = client.create_agenda(
    topic="Climate adaptation strategies",
    duration_days=90
)

# Verify a hypothesis
verification = client.verify(
    hypothesis="Solar geoengineering reduces crop yields",
    sources=["pubmed", "nature", "arxiv"]
)

MCP Tools

c4reqber exposes 21 Model Context Protocol tools for integration with Claude, Cursor, and other MCP-compatible agents.

Tool	Description
`c4_solve`	12-stage discovery pipeline
`c4_search`	Search 47 knowledge sources
`c4_triz`	TRIZ contradiction resolution
`c4_verify`	Formal proof verification (Lean4/Coq/Dafny/Agda/Z3/Hoare)
`c4_prove`	LLM-based hypothesis proving with iterative error correction
`c4_simulate`	Physics simulation via 38 engine adapters
`c4_bayesian`	Bayesian inference (MCMC/BMA)
`c4_causal`	Causal discovery (do-calculus)
`c4_export`	Export to Markdown/JSON/BibTeX/LaTeX
`c4_social`	Social publishing — preprint upload, ORCID, multi-platform posting
`c4_fingerprint`	C4 Z₃³ state classification
`c4_transfer`	Cross-domain isomorphism transfer
`c4_meta`	Meta-cognitive reflection
`c4_chain`	C4 discovery chain (Theorem 11)
`c4_codegen`	Code generation via MCP tool
`c4_autoresearch`	Karpathy-style ML training loop
`blast_solve`	UniversalSolvePipeline strategic artifacts
`blast_turbo`	HILDiscoveryPipeline paradigm dissertations
`blast_flash`	Quick LLM + USP cognitive analysis
`blast_turbofactory`	Parallel pipeline factory
`blast_auto`	Auto-router dispatch

Authentication

All API requests require a Bearer token in the Authorization header. Obtain your token from the GitLab repository or self-host.

Authorization: Bearer sk_c4r_xxxxxxxxxxxxxxxx

Rate Limits

Tier	Requests/min	Concurrent
Free	10	1
Pro	120	5
Enterprise	Unlimited	50