# Engraph

> Index and search document collections using hybrid semantic + graph + full-text search. Use when users need to search knowledge bases, find connections between documents, discover related content via link graphs, or query indexed markdown collections. Use when this capability is needed.

- Skill: `tomevault-io/engraph` (Agent Skill, multi-file: 2 files)
- Install (CLI): `npx skillmds@latest add tomevault-io/engraph`
- Raw SKILL.md: https://api.skillmd.com/api/skills/tomevault-io/engraph/raw
- Safety review: pending (external: skill-scanner PASS, skillspector PASS)
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Docs & Writing
- Author: tomevault-io (https://skillmd.com/u/tomevault-io)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/tomevault-io/engraph

---


# Engraph — Hybrid Semantic & Graph Search for Document Collections

Local knowledge engine for markdown document collections. Combines semantic embeddings, full-text search (BM25), wikilink graph traversal, temporal scoring, and cross-encoder reranking.

## Status

!`engraph --version 2>/dev/null || echo "Not installed: brew install devwhodevs/tap/engraph"`

## Indexing

```bash
engraph index /path/to/documents        # Incremental; only changed files re-embedded
engraph index /path/to/documents --rebuild
engraph status                          # File count, stats, index freshness
engraph clear                           # Drop the index (--all also removes models)
```

## Search

```bash
engraph search "how does the auth flow work"
engraph search "performance regressions last month" --explain
engraph search "architecture decisions" -n 5 --json
```

| Flag              | Description                                    |
| ----------------- | ---------------------------------------------- |
| `-n, --top-n <N>` | Number of results (default: from config or 10) |
| `--explain`       | Show per-lane RRF score breakdown              |
| `--json`          | Machine-readable JSON output                   |

### Query Tips

- **Conceptual / vague**: Use natural language. The orchestrator classifies intent and boosts semantic weight automatically.
- **Keyword-heavy**: Exact terms, identifiers, and names work well via the BM25 lane.
- **Temporal**: "last week", "yesterday", "March 2026" — the temporal lane activates automatically.

## Graph Inspection

```bash
engraph graph show "path/to/note.md"    # Connections for a document
engraph graph show "#docid"             # By document ID
engraph graph stats                     # Nodes, edges, density
```

## Context Queries

```bash
engraph context topic "authentication" --budget 8000
engraph context who "Person Name"
engraph context project "Project Name"
engraph context vault-map                 # Collection structure overview
engraph context read "path/to/note.md"    # Full content + metadata
engraph context list --tags architecture  # Filter by tags, folder, created_by, etc.
```

> Health diagnostics (orphans, broken links, stale notes, tag hygiene) are exposed through the MCP `health` tool and the HTTP `GET /api/health` endpoint — see `references/http-rest-api.md`.

## Setup

```bash
engraph index /path/to/documents
engraph search "your query"
```

## References

- `references/mcp-setup.md` — configure engraph as an MCP server (Claude Code, Claude Desktop).
- `references/http-rest-api.md` — HTTP REST API endpoints, authentication, and examples for web agents and scripts.

---
> Source: [devwhodevs/engraph](https://github.com/devwhodevs/engraph) — distributed by [TomeVault](https://tomevault.io).
<!-- tomevault:4.0:skill_md:2026-06-19 -->