rust-memex 0.6.4

Operator CLI + MCP server: canonical corpus second: semantic index second to aicx
Documentation
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
# rust-memex
[![Crates.io](https://img.shields.io/crates/v/rust-memex)](https://crates.io/crates/rust-memex) [![License](https://img.shields.io/crates/l/rust-memex)](LICENSE) [![Downloads](https://img.shields.io/crates/d/rust-memex)](https://crates.io/crates/rust-memex) [![CI](https://github.com/Loctree/rust-memex/actions/workflows/ci.yml/badge.svg)](https://github.com/Loctree/rust-memex/actions)

`rust-memex` is a custom Rust MCP kernel providing RAG and long-term memory capabilities to AI agents via LanceDB.

It exposes two explicit transport modes from a single canonical surface:
1.  **`stdio` (Standard MCP)**: Native MCP integration for local agents (e.g., Claude Desktop).
2.  **`HTTP/SSE` (Multi-Agent Daemon)**: A central daemon mode allowing concurrent AI agents to access the same memory pool over the network, resolving LanceDB's exclusive lock constraints.

> **Binary Name:** `rust-memex` is the only supported binary name. The GitHub installer also creates `rust_memex` as a legacy compatibility symlink for older scripts.
>
> **MCP Contract:** The current MCP surface is intentionally tools-only. `initialize` advertises `tools`, while `resources/*` is not implemented yet.

## Release Surface

- Quick install: `curl -LsSf https://raw.githubusercontent.com/Loctree/rust-memex/main/install.sh | sh`
- Prebuilt binary bundles: [GitHub Releases]https://github.com/Loctree/rust-memex/releases uploaded from locally built and signed artifacts
- Release runbook: [docs/RELEASE.md]docs/RELEASE.md
- Configuration guide: [docs/02_configuration.md]docs/02_configuration.md
- HTTP/SSE reference: [docs/HTTP_API.md]docs/HTTP_API.md
- Static launch page source: `docs/index.html` published by `.github/workflows/pages.yml`

## Quick Start

```bash
# Install from the latest GitHub Release
curl -LsSf https://raw.githubusercontent.com/Loctree/rust-memex/main/install.sh | sh

# Start the MCP server
rust-memex serve

# Open the local dashboard in your browser
rust-memex dashboard

# Or run the multi-agent HTTP/SSE daemon
rust-memex sse
```

## Overview

As an MCP (Model Context Protocol) server, `rust-memex` provides:
- **RAG (Retrieval-Augmented Generation)** - document indexing and semantic search
- **Hybrid Search** - BM25 keyword + semantic vector search (Tantivy-based)
- **Vector Memory** - semantic storage and retrieval of text chunks
- **Namespace Isolation** - data isolation in namespaces
- **Security** - token-based access control for protected namespaces
- **Onion Slice Architecture** - hierarchical embeddings (OUTER→MIDDLE→INNER→CORE)
- **Preprocessing** - automatic noise filtering from conversation exports (~36-40% reduction)
- **Exact-Match Deduplication** - SHA256-based dedup for overlapping exports

## Architecture

```
┌─────────────────────────────────────────────────────────────┐
│                      rust-memex                              │
├─────────────────────────────────────────────────────────────┤
│  MCP Server (JSON-RPC over stdio)                           │
│  ├── handlers/mod.rs    - Request routing & validation      │
│  ├── security/mod.rs    - Namespace access control          │
│  └── rag/mod.rs         - RAG pipeline                      │
├─────────────────────────────────────────────────────────────┤
│  Storage Layer                                               │
│  ├── LanceDB           - Vector embeddings                  │
│  ├── Tantivy           - BM25 keyword index                 │
│  └── moka              - In-memory cache                    │
├─────────────────────────────────────────────────────────────┤
│  Embeddings (External Providers)                             │
│  ├── Ollama            - Local models (recommended)         │
│  ├── MLX Bridge        - Apple Silicon acceleration         │
│  └── OpenAI-compatible - Any compatible endpoint            │
└─────────────────────────────────────────────────────────────┘
```

## Features

### RAG Tools
| Tool | Description |
|------|-------------|
| `rag_index` | Index document from file |
| `rag_index_text` | Index raw text |
| `rag_search` | Search documents semantically (supports `auto_route`) |

### Memory Tools
| Tool | Description |
|------|-------------|
| `memory_upsert` | Add/update chunk in namespace |
| `memory_get` | Get chunk by ID |
| `memory_search` | Search semantically in namespace (supports `auto_route`) |
| `memory_delete` | Delete chunk |
| `memory_purge_namespace` | Delete all chunks in namespace |
| `dive` | Deep exploration with all onion layers (outer/middle/inner/core) |

### Security Tools
| Tool | Description |
|------|-------------|
| `namespace_create_token` | Create access token for namespace |
| `namespace_revoke_token` | Revoke token (namespace becomes public) |
| `namespace_list_protected` | List protected namespaces |
| `namespace_security_status` | Security system status |

## Library Usage

`rust-memex` can be used as a library in your Rust applications. It provides a high-level `MemexEngine` API for vector storage operations.

### Add to Cargo.toml

```toml
# Full library with CLI
rust-memex = "0.5"

# Library only (no CLI dependencies)
rust-memex = { version = "0.5", default-features = false }
```

### Basic Usage

```rust
use rust_memex::{MemexEngine, MemexConfig, MetaFilter, StoreItem};
use serde_json::json;

#[tokio::main]
async fn main() -> anyhow::Result<()> {
    // Quick setup for any application
    let engine = MemexEngine::for_app("my-app", "documents").await?;

    // Store a document
    engine.store(
        "doc-1",
        "Patient presented with lethargy and decreased appetite",
        json!({"patient_id": "P-123", "visit_type": "checkup"})
    ).await?;

    // Search semantically
    let results = engine.search("lethargy symptoms", 10).await?;
    for r in &results {
        println!("{}: {} (score: {:.2})", r.id, r.text, r.score);
    }

    // Get by ID
    if let Some(doc) = engine.get("doc-1").await? {
        println!("Found: {}", doc.text);
    }

    // Delete
    engine.delete("doc-1").await?;

    Ok(())
}
```

### Vista Integration

For Vista PIMS, use the optimized constructor:

```rust
use rust_memex::MemexEngine;

// Vista-optimized: 1024 dims, qwen3-embedding:0.6b model
let engine = MemexEngine::for_vista().await?;

// Store visit notes
engine.store(
    "visit-456",
    "SOAP note: Feline diabetes mellitus diagnosis...",
    json!({"patient_id": "P-789", "doc_type": "soap_note"})
).await?;
```

### Batch Operations

```rust
use rust_memex::{MemexEngine, StoreItem};
use serde_json::json;

let engine = MemexEngine::for_app("my-app", "notes").await?;

let items = vec![
    StoreItem::new("doc-1", "First document").with_metadata(json!({"type": "note"})),
    StoreItem::new("doc-2", "Second document").with_metadata(json!({"type": "note"})),
    StoreItem::new("doc-3", "Third document").with_metadata(json!({"type": "note"})),
];

let result = engine.store_batch(items).await?;
println!("Stored {} documents", result.success_count);
```

### GDPR-Compliant Deletion

```rust
use rust_memex::{MemexEngine, MetaFilter};

let engine = MemexEngine::for_app("my-app", "patients").await?;

// Delete all documents for a specific patient
let filter = MetaFilter::for_patient("P-123");
let deleted = engine.delete_by_filter(filter).await?;
println!("Deleted {} documents", deleted);
```

### Hybrid Search (BM25 + Vector)

```rust
use rust_memex::{MemexEngine, SearchMode};

let engine = MemexEngine::for_app("my-app", "documents").await?;

// Hybrid search with BM25 + vector fusion (recommended)
let results = engine.search_hybrid("dragon mac studio", 10).await?;
for r in &results {
    println!("{}: {} (combined: {:.2}, vector: {:.2}, bm25: {:.2})",
        r.id, r.document, r.combined_score, r.vector_score, r.bm25_score);
}

// Explicit mode selection
let results = engine.search_with_mode("exact keyword", 10, SearchMode::Keyword).await?;
let results = engine.search_with_mode("semantic concept", 10, SearchMode::Vector).await?;
let results = engine.search_with_mode("best of both", 10, SearchMode::Hybrid).await?;
```

### MCP Contract and In-Process Helpers

Use `tool_definitions()` when you need the exact MCP tool metadata exposed by
the stdio and HTTP/SSE transports. The helper functions below are for in-process
Rust callers; they are a local convenience layer, not a second public MCP
contract. Their names intentionally differ from MCP tool names so the two
surfaces do not drift together by accident.

```rust
use rust_memex::{MemexEngine, search_documents, store_document, tool_definitions};
use serde_json::json;

let engine = MemexEngine::for_app("agent", "memory").await?;

// Canonical MCP tool surface exposed by rust-memex transports
let tools = tool_definitions();
assert!(tools.iter().any(|tool| tool.name == "memory_upsert"));

// Local Rust helpers for in-process callers
let result = store_document(
    &engine,
    "mem-1".to_string(),
    "Important information to remember".to_string(),
    json!({"source": "conversation"}),
).await?;
assert!(result.success);

let results = search_documents(&engine, "important".to_string(), 5, None).await?;
```

### Feature Flags

| Feature | Description | Default |
|---------|-------------|---------|
| `cli` | CLI binary, TUI wizard, progress bars | Yes |
| `provider-cascade` | Ollama/OpenAI-compatible embeddings | Yes |

```bash
# Build library only (no CLI)
cargo build --no-default-features

# Build with CLI
cargo build --features cli
```

---

## Configuration Guide

Complete guide for integrating rust-memex as a library in any Rust project.

### Prerequisites

**Ollama** (recommended) or any OpenAI-compatible embedding API:

```bash
# Install Ollama
curl -fsSL https://ollama.com/install.sh | sh

# Pull an embedding model (choose based on your needs)
ollama pull qwen3-embedding:0.6b    # 1024 dims, ~600MB (fast, good quality)
ollama pull qwen3-embedding:8b      # 4096 dims, ~4GB (best quality)
ollama pull nomic-embed-text        # 768 dims, ~274MB (lightweight)

# Verify it's running
curl http://localhost:11434/api/tags
```

### Environment Variables

Configure via `.env` or environment:

```bash
# =============================================================================
# EMBEDDING PROVIDER CONFIGURATION
# =============================================================================

# Ollama (default, recommended)
OLLAMA_BASE_URL=http://localhost:11434
EMBEDDING_MODEL=qwen3-embedding:0.6b
EMBEDDING_DIMENSION=1024

# Database storage (auto-created)
MEMEX_DB_PATH=~/.rmcp-servers/myapp/lancedb

# Optional: BM25 keyword search index
MEMEX_BM25_PATH=~/.rmcp-servers/myapp/bm25

# =============================================================================
# ADVANCED: Multiple providers (fallback cascade)
# =============================================================================

# Remote embedding server fallback
# DRAGON_BASE_URL=http://your-server.local
# DRAGON_EMBEDDER_PORT=12345

# MLX embedder for Apple Silicon
# EMBEDDER_PORT=12300
# MLX_MAX_BATCH_CHARS=32000
# MLX_MAX_BATCH_ITEMS=16
# DISABLE_MLX=1  # Set to disable MLX fallback
```

### Quick Start (Auto-config)

```rust
use rust_memex::MemexEngine;
use serde_json::json;

// Auto-configures from defaults + environment
let engine = MemexEngine::for_app("my-app", "default").await?;

engine.store("doc-1", "Document content...", json!({"type": "note"})).await?;
let results = engine.search("content", 10).await?;
```

### Custom Configuration

```rust
use rust_memex::{MemexConfig, MemexEngine};
use rust_memex::embeddings::{EmbeddingConfig, ProviderConfig};

// Read from your app's environment
let ollama_url = std::env::var("OLLAMA_BASE_URL")
    .unwrap_or_else(|_| "http://localhost:11434".to_string());
let model = std::env::var("EMBEDDING_MODEL")
    .unwrap_or_else(|_| "qwen3-embedding:0.6b".to_string());
let dimension: usize = std::env::var("EMBEDDING_DIMENSION")
    .unwrap_or_else(|_| "1024".to_string())
    .parse()
    .unwrap_or(1024);
let db_path = std::env::var("MEMEX_DB_PATH")
    .unwrap_or_else(|_| "~/.rmcp-servers/myapp/lancedb".to_string());

let config = MemexConfig {
    app_name: "my-app".to_string(),
    namespace: "default".to_string(),
    db_path: Some(db_path),
    dimension,
    embedding_config: EmbeddingConfig {
        required_dimension: dimension,
        providers: vec![ProviderConfig {
            name: "ollama".to_string(),
            base_url: ollama_url,
            model,
            priority: 1,
            endpoint: "/v1/embeddings".to_string(),
        }],
        ..EmbeddingConfig::default()
    },
    enable_bm25: false,
    bm25_config: None,
};

let engine = MemexEngine::new(config).await?;
```

### Provider Cascade (Multiple Fallbacks)

Configure multiple providers - library tries them in priority order:

```rust
use rust_memex::embeddings::{EmbeddingConfig, ProviderConfig};

let config = EmbeddingConfig {
    required_dimension: 1024,
    providers: vec![
        // Priority 1: Local Ollama (fastest)
        ProviderConfig {
            name: "ollama-local".to_string(),
            base_url: "http://localhost:11434".to_string(),
            model: "qwen3-embedding:0.6b".to_string(),
            priority: 1,
            endpoint: "/v1/embeddings".to_string(),
        },
        // Priority 2: Remote server fallback
        ProviderConfig {
            name: "remote-server".to_string(),
            base_url: "http://your-server:8080".to_string(),
            model: "text-embedding-3-small".to_string(),
            priority: 2,
            endpoint: "/v1/embeddings".to_string(),
        },
        // Priority 3: OpenAI API fallback
        ProviderConfig {
            name: "openai".to_string(),
            base_url: "https://api.openai.com".to_string(),
            model: "text-embedding-3-small".to_string(),
            priority: 3,
            endpoint: "/v1/embeddings".to_string(),
        },
    ],
    ..EmbeddingConfig::default()
};
```

### Namespace Strategy

**Recommended: One namespace per application, use metadata for filtering:**

```rust
// ✅ CORRECT: Single namespace, filter by user_id/entity_id in metadata
let engine = MemexEngine::for_app("my-app", "default").await?;

// Store with entity IDs in metadata
engine.store("doc-1", "Document content...", json!({
    "user_id": "U-123",      // For multi-tenant filtering
    "project_id": "P-456",   // For project-level filtering
    "doc_type": "note",
    "created_at": "2024-12-28"
})).await?;

// Search within user context
let filter = MetaFilter::default().with_custom("user_id", "U-123");
let results = engine.search_filtered("query", filter, 10).await?;

// GDPR deletion: remove all user data
let deleted = engine.delete_by_filter(
    MetaFilter::default().with_custom("user_id", "U-123")
).await?;
```

### Embedding Models Reference

| Model | Dimensions | Size | Use Case |
|-------|------------|------|----------|
| `qwen3-embedding:0.6b` | 1024 | ~600MB | Fast, good quality (recommended) |
| `qwen3-embedding:8b` | 4096 | ~4GB | Best quality, slower |
| `nomic-embed-text` | 768 | ~274MB | Lightweight, fast |
| `mxbai-embed-large` | 1024 | ~670MB | Good multilingual |
| `all-minilm` | 384 | ~46MB | Very fast, lower quality |

### Troubleshooting

**Error: "No embedding providers available"**
```bash
# Check if Ollama is running
curl http://localhost:11434/api/tags

# Start Ollama
ollama serve

# Pull model if missing
ollama pull qwen3-embedding:0.6b
```

**Error: "Dimension mismatch"**
- LanceDB dimension is fixed per table after creation
- Use different `db_path` for different dimensions
- Delete old database to change dimensions

**Error: "Connection refused"**
```bash
# Linux
systemctl status ollama
systemctl start ollama

# macOS
brew services info ollama
brew services start ollama

# Or run manually
ollama serve
```

**Performance tuning:**
```bash
# Larger batches (requires more VRAM)
MLX_MAX_BATCH_CHARS=64000
MLX_MAX_BATCH_ITEMS=32
```

---

## Quick Start

### Installation

**Quick install (recommended):**
```bash
curl -LsSf https://raw.githubusercontent.com/Loctree/rust-memex/main/install.sh | sh
```

Prebuilt GitHub Release bundles are the canonical install path and avoid compiling
the full LanceDB-heavy Rust dependency graph on the target machine.

**From source (development only):**
```bash
cargo install --path .
```

### Running

```bash
# Canonical MCP surface
rust-memex serve

# With security enabled
rust-memex serve --security-enabled

# With HTTP/SSE server for multi-agent access
rust-memex serve --http-port 8997

# HTTP-only daemon mode (no MCP stdio)
rust-memex serve --http-port 8997 --http-only
```

`rust-memex` always exposes one canonical MCP tool surface. To narrow runtime
access, use `--allowed-paths`, HTTP auth, or namespace security rather than a
separate "memory/full" mode switch.

## HTTP/SSE Server (Multi-Agent Access)

LanceDB uses exclusive file locks - only one process can access the database at a time.
The HTTP/SSE server solves this by providing a central access point for multiple agents.

### Architecture

```
┌─────────────────────────────────────────────────────────────┐
│                     rust-memex daemon                        │
│  ┌─────────────────┐    ┌─────────────────┐                 │
│  │   MCP Server    │    │   HTTP/SSE      │                 │
│  │   (stdio)       │    │   (port 8997)   │                 │
│  └────────┬────────┘    └────────┬────────┘                 │
│           │                      │                          │
│           └──────────┬───────────┘                          │
│                      ▼                                      │
│              ┌─────────────┐                                │
│              │ RAGPipeline │ ← Single lock holder           │
│              └──────┬──────┘                                │
│                     ▼                                       │
│              ┌─────────────┐                                │
│              │   LanceDB   │                                │
│              └─────────────┘                                │
└─────────────────────────────────────────────────────────────┘
         ▲                    ▲
         │                    │
    Claude Desktop       HTTP Agents
    (MCP stdio)          (curl, fetch)
```

### HTTP Endpoints

| Endpoint | Method | Description |
|----------|--------|-------------|
| `/health` | GET | Health check (status, db_path, embedding_provider) |
| `/search` | POST | Search with optional `project`, `layer`, and `deep` filters (`k` alias supported); response includes collapsed `clusters` and `duplicate_count` |
| `/sse/search` | GET | SSE streaming search with optional `project`, `layer`, and `deep` filters |
| `/api/context-pack` | POST | Build a markdown context pack from a query or explicit chunk IDs, with grouped evidence and rebuilt indexed source chunks |
| `/upsert` | POST | Add/update document |
| `/index` | POST | Full pipeline indexing with onion slices |
| `/expand/{ns}/{id}` | GET | Expand onion slice (get children) |
| `/parent/{ns}/{id}` | GET | Drill up to parent slice |
| `/get/{ns}/{id}` | GET | Get document by ID |
| `/delete/{ns}/{id}` | POST | Delete document |
| `/ns/{namespace}` | DELETE | Purge entire namespace |

**Diagnostic & lifecycle endpoints** (require `Bearer` auth_token):

| Endpoint | Method | Description |
|----------|--------|-------------|
| `/api/audit` | GET | Per-namespace quality audit (chunk completeness, hash coverage, score) |
| `/api/stats` / `/api/stats/{ns}` | GET | Database / namespace statistics |
| `/api/timeline` | GET | Indexing timeline aggregates |
| `/api/purge-quality` | POST | Purge low-quality chunks under a threshold (gated by approval key) |
| `/api/dedup` | POST | Run post-index deduplication (`group-by`, `keep`, `dry_run` body fields) |
| `/api/backfill-hashes` | POST | Spec P0 backfill: populate per-chunk `content_hash` + `source_hash` for pre-v4 namespaces |

### MCP-over-SSE Endpoints (Claude Code compatibility)

| Endpoint | Method | Description |
|----------|--------|-------------|
| `/sse/` | GET | SSE stream - sends `endpoint` event with messages URL |
| `/messages/` | POST | JSON-RPC messages with `?session_id=xxx` |

Configure in `~/.claude.json`:
```json
{
  "mcpServers": {
    "rust-memex": {
      "type": "sse",
      "url": "http://localhost:8997/sse/"
    }
  }
}
```

### Usage Examples

```bash
# Open the local dashboard
rust-memex dashboard --db-path ~/.ai-memories/lancedb

# Start daemon
rust-memex sse --db-path ~/.ai-memories/lancedb &

# Health check
curl http://localhost:8997/health

# Store document
curl -X POST http://localhost:8997/upsert \
  -H "Content-Type: application/json" \
  -d '{"namespace": "agent1", "id": "mem1", "content": "Important context..."}'

# Search
curl -X POST http://localhost:8997/search \
  -H "Content-Type: application/json" \
  -d '{"query": "context", "namespace": "agent1", "k": 10, "project": "Vista", "deep": true}'

# SSE streaming search
curl -N "http://localhost:8997/sse/search?query=context&namespace=agent1&limit=5&project=Vista&layer=1"
```

### Multi-Host Database Paths

For setups with multiple machines (e.g., dragon, mgbook16), use per-host database paths:

```bash
# Per-host paths (each machine gets own database)
rust-memex serve --db-path ~/.ai-memories/lancedb.$(hostname -s)

# Or use the wizard for machine-agnostic configuration
rust-memex wizard
```

The TUI wizard auto-detects hostname and offers:
- **Shared mode**: `~/.ai-memories/lancedb` (same path everywhere)
- **Per-host mode**: `~/.ai-memories/lancedb.dragon`, `~/.ai-memories/lancedb.mgbook16`, etc.

### Configuration (TOML)

```toml
# ~/.rmcp-servers/rust-memex/config.toml

db_path = "~/.rmcp-servers/rust-memex/lancedb"
cache_mb = 4096
log_level = "info"

# Whitelist of allowed paths
allowed_paths = [
    "~",
    "/Volumes/ExternalDrive/data"
]

# Security
security_enabled = true
token_store_path = "~/.rmcp-servers/rust-memex/tokens.json"

# Optional: dashboard-only OIDC for browser users.
# API / SSE / MCP still stay Bearer-authenticated via auth_token.
auth_token = "replace-me"

[dashboard_oidc]
issuer_url = "https://issuer.example"
client_id = "rust-memex-dashboard"
client_secret = "optional-confidential-client-secret"
public_base_url = "https://memex.example.com"
scopes = ["openid", "profile", "email"]
```

## Documentation

- [01_security.md]./01_security.md - Security system (namespace tokens)
- [02_configuration.md]./02_configuration.md - Configuration and CLI options

## Onion Slice Architecture

Instead of traditional flat chunking, rust-memex offers hierarchical "onion slices":

```
┌─────────────────────────────────────────┐
│  OUTER (~100 chars)                     │  ← Minimum context, maximum navigation
│  Keywords + ultra-compression           │
├─────────────────────────────────────────┤
│  MIDDLE (~300 chars)                    │  ← Key sentences + context
├─────────────────────────────────────────┤
│  INNER (~600 chars)                     │  ← Expanded content
├─────────────────────────────────────────┤
│  CORE (full text)                       │  ← Complete document
└─────────────────────────────────────────┘
```

**Philosophy:** "Minimum info → Maximum navigation paths"

### QueryRouter & Auto-Route

Intelligent query intent detection for automatic search mode selection:

```bash
# Auto-detect query intent and select optimal mode
rust-memex search -n memories -q "when did we buy dragon" --auto-route
# Output: Query intent: temporal (confidence: 0.70)
#         Selects: hybrid mode with date boosting

# Structural queries suggest loctree
rust-memex search -n code -q "who imports main.rs" --auto-route
# Output: Query intent: structural (confidence: 0.80)
#         Consider: loctree query --kind who-imports --target main.rs

# Deep exploration with all onion layers
rust-memex dive -n memories -q "dragon" --verbose
```

**Intent Types:**
| Intent | Trigger Keywords | Recommended Mode |
|--------|-----------------|------------------|
| Temporal | when, date, yesterday, ago, 2024 | Hybrid (date boost) |
| Structural | import, depends, module, who uses | BM25 + loctree suggestion |
| Semantic | similar, related, explain | Vector |
| Exact | "quoted strings" | BM25 |
| Hybrid | (default) | Vector + BM25 fusion |

### CLI Commands

```bash
# Index with onion slicing (default)
rust-memex index -n memories /path/to/data/ --slice-mode onion

# Index with progress bar and ETA
rust-memex index -n memories /path/to/data/ --progress

# Index with flat chunking (backward compatible)
rust-memex index -n memories /path/to/data/ --slice-mode flat

# Search in namespace
rust-memex search -n memories -q "best moments" --limit 10

# Search only in specific layer
rust-memex search -n memories -q "query" --layer outer

# Drill down in hierarchy (expand children)
rust-memex expand -n memories -i "slice_id_here"

# Get chunk by ID
rust-memex get -n memories -i "chunk_abc123"

# RAG search (cross-namespace)
rust-memex rag-search -q "search term" --limit 5

# List namespaces with stats
rust-memex namespaces --stats

# Export namespace to JSON
rust-memex export -n memories -o backup.json --include-embeddings
```

### Preprocessing (Noise Filtering)

Automatic removal of ~36-40% noise from conversation exports:
- MCP tool artifacts (`<function_calls>`, `<invoke>`, etc.)
- CLI output (git status, cargo build, npm install)
- Metadata (UUIDs, timestamps → placeholders)
- Empty/boilerplate content

```bash
# Index with preprocessing
rust-memex index -n memories /path/to/export.json --preprocess
```

### Deduplication & Hash Hygiene

Two layers of dedup work together:

**1. Pre-index source dedup** (during `index`)

The pipeline computes `sha256(file_text)` and skips files whose `source_hash`
already exists in the namespace (with a fallback to `content_hash` for pre-v4
namespaces). The skip line is logged at `info!` so it shows up in the default
operator run log:

```
Skip duplicate source: /path/to/file.md (source_hash 8ee43c1e7393b432)
```

```bash
# Dedup enabled (default)
rust-memex index -n memories /path/to/data/

# Disable dedup for this run
rust-memex index -n memories /path/to/data/ --no-dedup

# Spec P4 escape hatch: force re-index a known-duplicate source
rust-memex index -n memories /path/to/data/ --allow-duplicates
```

**2. Post-index dedup CLI** (standalone command)

```bash
# Default grouping: source-hash + layer (preserves onion structure,
# removes only true source repeats while keeping outer/middle/inner/core)
rust-memex dedup -n kb:transcripts --dry-run

# Collapse all layers per source (legacy aggressive grouping)
rust-memex dedup -n kb:transcripts --group-by source-hash

# Per-chunk content_hash grouping (legacy pre-v4 behavior)
rust-memex dedup -n kb:transcripts --group-by content-hash

# Cross-namespace dedup pool
rust-memex dedup --cross-namespace --dry-run false

# Keep newest duplicates instead of oldest
rust-memex dedup -n memories --keep newest
```

**3. Hash backfill (spec P0)** — fills `content_hash` (per-chunk) and
`source_hash` (per-source) for namespaces indexed before v4. Without backfill,
`dedup` reports "Without hash: N (cannot deduplicate)" and is blind to legacy
chunks:

```bash
# Dry-run backfill across all namespaces
rust-memex backfill-hashes

# Backfill one namespace, then commit
rust-memex backfill-hashes -n kb:transcripts --dry-run false

# Machine-readable JSON for scripts / CI
rust-memex backfill-hashes --json
```

**Output with statistics:**
```
Indexing complete:
  New chunks:          234
  Files indexed:       67
  Skipped (duplicate): 33
  Deduplication:       enabled
```

### LLM-Synthesized Outer Layer (Spec P3)

The default outer layer is a TF-based keyword extract (`--outer-synthesis
keyword`). For transcript-heavy namespaces where keyword splat is noise (CLI
animation gerunds, structural markdown, file-path tokens), the outer layer can
be replaced with a 1-3 sentence summary from a local Ollama model:

```bash
# Wire the outer layer through Ollama (requires --pipeline mode)
rust-memex index -n kb:transcripts /path/to/transcripts/ \
  --slice-mode onion \
  --pipeline \
  --outer-synthesis llm \
  --ollama-model qwen2.5:3b \
  --ollama-endpoint http://localhost:11434
```

Failure modes (network, non-2xx, malformed JSON, empty completion) silently
fall back to the keyword outer so the pipeline never stalls. Reachable only
through `--pipeline` — passing `--outer-synthesis llm` without `--pipeline` is
rejected up-front by clap.

## Code Structure

```
rust-memex/
├── src/
│   ├── lib.rs              # Public API & ServerConfig
│   ├── bin/
│   │   └── rust-memex.rs   # CLI binary (serve, index, search, get, expand, etc.)
│   ├── handlers/
│   │   └── mod.rs          # MCP request handlers
│   ├── security/
│   │   └── mod.rs          # Namespace access control
│   ├── rag/
│   │   └── mod.rs          # RAG pipeline + OnionSlice architecture
│   ├── preprocessing/
│   │   └── mod.rs          # Noise filtering for conversation exports
│   ├── storage/
│   │   └── mod.rs          # LanceDB + Tantivy (schema v4: source_hash + per-chunk content_hash)
│   ├── embeddings/
│   │   └── mod.rs          # MLX/FastEmbed bridge
│   └── tui/
│       └── mod.rs          # Configuration wizard
└── Cargo.toml
```

## Claude/MCP Integration

Add to `~/.claude.json`:

```json
{
  "mcpServers": {
    "rust-memex": {
      "command": "rust-memex",
      "args": ["serve", "--security-enabled"]
    }
  }
}
```

---

Vibecrafted with AI Agents by Loctree (c)2025 The LibraxisAI Team
Co-Authored-By: [Maciej](void@div0.space) & [Klaudiusz](the1st@whoai.am)