Getting Started
Install c4reqber, run your first discovery pipeline, and configure MCP tools. From zero to paradigm shift in 5 minutes.
1. Installation
pip install c4reqber
Package: pypi.org/project/c4reqber
Requires Python 3.11+.
Verify the installation:
blast --help # c4reqber v5.7.0
Packages vs API keys (different commands):
blast setup— scientific packages (GROMACS, OpenMM, …)blast initorblast config keys— API keys →~/.c4reqber/secrets.env
2. Quick Commands
| Command | What it does |
|---|---|
blast solve "problem" | One-shot discovery pipeline → article/blueprint/whitepaper |
blast turbo "topic" | Paradigm-shifting research proposal + verification |
blast flash "question" | Quick answer (2–10s) |
blast auto "query" | Auto-routed to best mode |
blast turbofactory "domain" | Parallel pipeline factory (mini/standard/mega/giga) |
blast tui | TUI v9 Cockpit — feed-driven discovery UI |
blast tui --demo --story=crispr | Demo mode without backend |
blast serve --mcp | MCP server for AI agents (21 tools) |
blast config keys | blast config keys — manage ~/.c4reqber/secrets.env |
blast social publish | blast social publish / blast social post |
TUI v9 overlays
Launch blast tui. Key overlays (also in : palette):
3. Your First Discovery
# One-shot discovery → strategic artifact
blast solve "quantum effects in photosynthesis"
# Paradigm-shifting research proposal
blast turbo "language evolution through horizontal gene transfer"
# Quick answer with sources
blast flash "What is the TRIZ principle of segmentation?" --sources
# TUI v9 Cockpit
blast tui
4. MCP Server Setup
c4reqber can be used as an MCP tool for AI agents (Claude, Cursor, Cline, Continue.dev). Add it to your MCP configuration:
// mcp-config.json
{
"mcpServers": {
"c4reqber": {
"command": "blast",
"args": ["serve", "--mcp"]
}
}
}
Available MCP Tools
| Tool | Description |
|---|---|
c4_solve | 12-stage discovery pipeline |
c4_search | 47 knowledge sources |
c4_triz | 40 TRIZ principles + CDI resolution |
c4_verify | Lean4/Coq/Dafny/Agda/Hoare/Z3 |
c4_simulate | 38 engine adapters |
c4_bayesian | MCMC / BMA inference |
c4_causal | Do-calculus discovery |
c4_fingerprint | Z₃³ state classification |
c4_export | LaTeX / Markdown / BibTeX |
c4_transfer | Cross-domain isomorphism |
c4_autoresearch | Karpathy-style ML loop |
c4_meta | Meta-cognitive reflection |
5. Configuration
# View current setup
blast config --show
blast config --show --json # TUI models overlay (Ctrl+Shift+M)
blast models --json
# Set model per phase
blast config --set F=claude-sonnet-4.6 --save
# API keys (primary store)
blast config keys
blast config keys --assign OPENROUTER_API_KEY=sk-or-...
# Set cost tier
blast config --tier balanced --save # $0.35/run
blast config --tier budget --save # $0.002/run
# Custom config dir (optional)
export C4REQBER_CONFIG=/path/to/config
6. Next Steps
- Social publishing guide — Zenodo, ORCID, Mastodon, Bluesky, …
- Register API keys — OpenRouter + scientific knowledge sources
- Configure GPU acceleration — Apple Silicon or NVIDIA CUDA
- Read the Architecture guide — understand 7-layer system design
- Explore the API reference — CLI, Python, REST, MCP
- See verified proposals — real outputs from the pipeline