Command-driven walkthroughs matched to how people actually use NeuralMind. Copy, run, done.
Canonical pages live at docs/use-cases/ — this wiki page is a jump table.
| Use case | Best for you if… | Primary outcome |
|---|---|---|
| Claude Code user | You use Claude Code daily | Two-phase optimization: retrieval + PostToolUse compression |
| Cost optimization | You need to reduce and report LLM spend | Baseline → measure → stakeholder-ready report |
| Any LLM (ChatGPT / Gemini / local) | You use a non-MCP chat or mixed models | Copy-paste + CLI-piped context into any model |
| Offline / regulated work | You’re on air-gapped or regulated systems | 100% local, zero telemetry, compliance properties table |
| Growing monorepo | Your codebase grows fast and drifts often | Three freshness strategies + large-repo tuning |
| If your goal is… | Do this | Expected outcome |
|---|---|---|
| Cut LLM spend on code Q&A | install-hooks + use query for questions |
5–10× total reduction vs baseline agent |
| Faster, more grounded agent responses | wakeup at session start → query / skeleton during |
Fewer hallucinations; less re-exploration |
| Keep all code local | Default install — no extra config | 100% offline; nothing leaves the machine |
| Work across Claude + GPT + Gemini | Build once, pipe output into any model | Model-agnostic |
| Make retrieval adapt to your team’s patterns | Enable memory + neuralmind learn . |
Reranking improves over time |
| Measure savings for a stakeholder | neuralmind benchmark . --json |
Per-query tokens, reduction ratios |
| Auto-refresh the index | neuralmind init-hook . |
Every commit rebuilds incrementally |
Full symptom/fix table lives in the main README.
You probably don’t need NeuralMind if:
You almost certainly want NeuralMind if:
See the Setup Guide once you’ve picked a use case.