@ruvector/rvf

Unified TypeScript/JavaScript SDK for the RuVector Format (RVF) — a cognitive container that stores vectors, carries models, boots compute kernels, and proves everything in a single .rvf file.

Platform Support

Deno: The WASM build targets wasm32-unknown-unknown, which runs natively in Deno. Import via npm: specifier or load the .wasm bundle directly.

Browser: The @ruvector/rvf-wasm package provides a ~46 KB control-plane WASM module plus a ~5.5 KB tile-compute module. Works in any browser with WebAssembly support.

Install

# Node.js (auto-detects native or WASM)
npm install @ruvector/rvf

# WASM only (browser, Deno, edge)
npm install @ruvector/rvf-wasm

Quick Start

Node.js

import { RvfDatabase } from '@ruvector/rvf';

// Create a vector store
const db = RvfDatabase.create('vectors.rvf', { dimension: 384 });

// Insert vectors
db.ingestBatch(new Float32Array(384), [1]);

// Query nearest neighbors
const results = db.query(new Float32Array(384), 10);

// Lineage & inspection
console.log(db.fileId());       // unique file UUID
console.log(db.dimension());    // 384
console.log(db.segments());     // [{ type, id, size }]

// Derive child (COW branching)
const child = db.derive('child.rvf');

db.close();

Browser (WASM)

<script type="module">
import init, { RvfStore } from '@ruvector/rvf-wasm';

await init();

const store = RvfStore.create(384, 'cosine');
store.ingest(new Float32Array(384), 0);
const results = store.query(new Float32Array(384), 10);
console.log('Results:', results);
</script>

Deno

// Import via npm: specifier
import init, { RvfStore } from "npm:@ruvector/rvf-wasm";

await init();

const store = RvfStore.create(384, 'cosine');
store.ingest(new Float32Array(384), 0);
const results = store.query(new Float32Array(384), 10);
console.log('Results:', results);

What is RVF?

RVF (RuVector Format) is a universal binary substrate that merges database, model, graph engine, kernel, and attestation into a single deployable file. A .rvf file is segmented — each segment carries a different payload type, and unknown segments are preserved by all tools.

Segment Types

N-API Methods (Node.js)

19 methods on the RvfDatabase class:

WASM Exports

29 exported functions for browser and edge runtimes:

Control plane (10): rvf_create, rvf_open, rvf_close, rvf_ingest, rvf_query, rvf_delete, rvf_status, rvf_compact, rvf_derive, rvf_segments

Tile compute (14): tile_dot_f32, tile_cosine_f32, tile_l2_f32, tile_dot_f16, tile_cosine_f16, tile_l2_f16, tile_topk, tile_quantize_sq8, tile_dequantize_sq8, tile_scan_filtered, tile_merge_topk, tile_batch_distance, tile_prefetch, tile_accumulate

Segment parsing (3): parse_segment_header, parse_vec_header, parse_manifest

Memory (2): rvf_alloc, rvf_free

CLI (Rust)

18 subcommands available through the rvf binary:

# Core operations
rvf create vectors.rvf --dimension 384 --metric cosine
rvf ingest vectors.rvf --input data.json
rvf query vectors.rvf --vector "[0.1,0.2,...]" --k 10
rvf delete vectors.rvf --ids "[1,2,3]"
rvf status vectors.rvf
rvf inspect vectors.rvf
rvf compact vectors.rvf

# Branching & lineage
rvf derive vectors.rvf --output child.rvf
rvf filter vectors.rvf --include "[1,2,3]"
rvf freeze vectors.rvf
rvf rebuild-refcounts vectors.rvf

# Compute containers
rvf serve vectors.rvf --port 8080
rvf launch vectors.rvf
rvf embed-kernel vectors.rvf --image bzImage
rvf embed-ebpf vectors.rvf --program filter.o

# Verification
rvf verify-witness vectors.rvf
rvf verify-attestation vectors.rvf

# Export
rvf export vectors.rvf --output dump.json

Build the CLI:

cargo install --path crates/rvf/rvf-cli

Example .rvf Files

45 pre-built example files are available for download (~11 MB total). These demonstrate every segment type and use case.

Download

# Download a specific example
curl -LO https://raw.githubusercontent.com/ruvnet/ruvector/main/examples/rvf/output/basic_store.rvf

# Clone just the examples
git clone --depth 1 --filter=blob:none --sparse https://github.com/ruvnet/ruvector.git
cd ruvector && git sparse-checkout set examples/rvf/output

Example Catalog

Generate Examples Locally

cd crates/rvf
cargo run --example generate_all
ls output/  # 45 .rvf files

Integration

With ruvector (npx ruvector)

The ruvector npm package includes 8 RVF CLI commands:

npm install ruvector @ruvector/rvf

# Enable RVF backend
export RUVECTOR_BACKEND=rvf

# Or use --backend flag
npx ruvector --backend rvf create mydb.rvf -d 384

# RVF-specific commands
npx ruvector rvf create mydb.rvf -d 384
npx ruvector rvf ingest mydb.rvf --input data.json
npx ruvector rvf query mydb.rvf --vector "[0.1,...]" --k 10
npx ruvector rvf status mydb.rvf
npx ruvector rvf segments mydb.rvf
npx ruvector rvf derive mydb.rvf --output child.rvf
npx ruvector rvf compact mydb.rvf
npx ruvector rvf export mydb.rvf --output dump.json

With rvlite

npm install rvlite @ruvector/rvf-wasm

When @ruvector/rvf-wasm is installed, rvlite can use RVF as a persistent storage backend:

import { createRvLite } from 'rvlite';

// rvlite auto-detects @ruvector/rvf-wasm for persistence
const db = await createRvLite({ dimensions: 384 });
await db.insert([0.1, 0.2, ...], { text: "Hello world" });
const results = await db.search([0.1, 0.2, ...], 5);

Packages

Crate Structure (Rust)

Real-World Examples

Self-Booting Microservice (Rust)

Create a single .rvf file that contains 50 vectors AND a bootable kernel — drop it on a VM and it boots:

use rvf_runtime::{RvfStore, RvfOptions, QueryOptions};
use rvf_runtime::options::DistanceMetric;
use rvf_types::kernel::{KernelArch, KernelType};

// 1. Create store with vectors
let mut store = RvfStore::create("bootable.rvf", RvfOptions {
    dimension: 128, metric: DistanceMetric::L2, ..Default::default()
})?;
store.ingest_batch(&vectors, &ids, None)?;

// 2. Embed a kernel — file now boots as a microservice
store.embed_kernel(
    KernelArch::X86_64 as u8,
    KernelType::Hermit as u8,
    0x0018,  // HAS_QUERY_API | HAS_NETWORKING
    &kernel_image,
    8080,
    Some("console=ttyS0 quiet"),
)?;

// 3. Verify everything is in one file
let (header, image) = store.extract_kernel()?.unwrap();
println!("Kernel: {} bytes, vectors: {}", image.len(), store.query(&q, 5, &QueryOptions::default())?.len());
store.close()?;
// Result: 31 KB file with vectors + kernel + witness chain

Run: cd examples/rvf && cargo run --example self_booting

Linux Microkernel Distribution

A single .rvf file as an immutable, bootable Linux distribution:

use rvf_runtime::{RvfStore, RvfOptions, MetadataEntry, MetadataValue, FilterExpr, QueryOptions};
use rvf_crypto::{create_witness_chain, sign_segment, verify_segment, shake256_256, WitnessEntry};
use ed25519_dalek::SigningKey;

// 1. Create system image with 20 packages as vector embeddings
let mut store = RvfStore::create("microkernel.rvf", options)?;
for pkg in packages {
    store.ingest_batch(&[&pkg.embedding], &[pkg.id], Some(&[MetadataEntry {
        key: "package".into(),
        value: MetadataValue::String(format!("{}@{}", pkg.name, pkg.version)),
    }]))?;
}

// 2. Embed kernel + SSH keys
store.embed_kernel(KernelArch::X86_64 as u8, KernelType::Linux as u8, 0x001F, &kernel, 8080, None)?;

// 3. Sign with Ed25519 — prevents unauthorized modifications
let signature = sign_segment(&segment_bytes, &signing_key);
verify_segment(&segment_bytes, &signature, &verifying_key)?;

// 4. Witness chain — every package install is audited
let chain = create_witness_chain(&witness_entries);
// Result: 14 KB file = bootable OS + packages + SSH + crypto + witness

Run: cd examples/rvf && cargo run --example linux_microkernel

Claude Code Appliance

Build an AI development environment as a single sealed file:

// Creates a .rvf file containing:
// - 20 development packages (rust, node, python, etc.)
// - Real Linux kernel with SSH on port 2222
// - eBPF XDP program for fast-path vector lookups
// - Vector store with development context embeddings
// - 6-entry witness chain for audit
// - Ed25519 + ML-DSA-65 signatures
let store = RvfStore::create("claude_code_appliance.rvf", options)?;
// ... embed packages, kernel, eBPF, witness chain, signatures ...
// Result: 5.1 MB sealed cognitive container

Run: cd examples/rvf && cargo run --example claude_code_appliance

Final file: 5.1 MB single .rvf — boots Linux, serves queries, runs Claude Code.

CLI Proof-of-Operations

# Full lifecycle in one session:

# Create a vector store
rvf create demo.rvf --dimension 128

# Ingest 100 vectors from JSON
rvf ingest demo.rvf --input data.json --format json

# Query nearest neighbors
rvf query demo.rvf --vector "0.1,0.2,0.3,..." --k 5

# Derive a COW child (only stores differences)
rvf derive demo.rvf child.rvf --type filter

# Inspect all segments
rvf inspect demo.rvf
# Output: MANIFEST_SEG (4 KB), VEC_SEG (51 KB), INDEX_SEG (12 KB)

# Verify witness chain integrity
rvf verify-witness demo.rvf

# Embed a kernel — file becomes self-booting
rvf embed-kernel demo.rvf --image bzImage --arch x86_64

# Launch in QEMU microVM
rvf launch demo.rvf --port 8080

# Compact and reclaim space
rvf compact demo.rvf

Witness Chain Verification

use rvf_crypto::{create_witness_chain, verify_witness_chain, shake256_256, WitnessEntry};

// Every operation is recorded in a tamper-evident hash chain
let entries = vec![
    WitnessEntry {
        prev_hash: [0; 32],
        action_hash: shake256_256(b"ingest: 1000 vectors, dim 384"),
        timestamp_ns: 1_700_000_000_000_000_000,
        witness_type: 0x01, // PROVENANCE
    },
    WitnessEntry {
        prev_hash: [0; 32], // linked by create_witness_chain
        action_hash: shake256_256(b"query: top-10, cosine"),
        timestamp_ns: 1_700_000_001_000_000_000,
        witness_type: 0x03, // SEARCH
    },
    WitnessEntry {
        prev_hash: [0; 32],
        action_hash: shake256_256(b"embed: kernel x86_64, 8080"),
        timestamp_ns: 1_700_000_002_000_000_000,
        witness_type: 0x02, // COMPUTATION
    },
];

let chain_bytes = create_witness_chain(&entries);
let verified = verify_witness_chain(&chain_bytes)?;
assert_eq!(verified.len(), 3);
// Changing any byte in any entry breaks the entire chain

COW Branching (Git-like for Vectors)

use rvf_runtime::{RvfStore, RvfOptions};
use rvf_types::DerivationType;

// Parent: 1M vectors (~512 MB)
let parent = RvfStore::create("parent.rvf", options)?;
parent.ingest_batch(&million_vectors, &ids, None)?;

// Child: shares all parent data, only stores changes
let child = parent.derive("child.rvf", DerivationType::Filter, None)?;
assert_eq!(child.lineage_depth(), 1);

// Modify 100 vectors → only 10 clusters copied (~2.5 MB, not 512 MB)
child.ingest_batch(&updated_vectors, &updated_ids, None)?;

// Query child — transparent parent resolution
let results = child.query(&query, 10, &QueryOptions::default())?;
// Results come from both local (modified) and inherited (parent) clusters

Generate All 45 Example Files

cd examples/rvf
cargo run --example generate_all
ls output/
# 45 .rvf files ready to inspect:
#   basic_store.rvf (152 KB)        — 1,000 vectors
#   self_booting.rvf (31 KB)        — vectors + kernel
#   linux_microkernel.rvf (15 KB)   — bootable OS image
#   claude_code_appliance.rvf (17 KB) — AI dev environment
#   sealed_engine.rvf (208 KB)      — signed inference engine
#   agent_memory.rvf (32 KB)        — AI agent memory
#   ... and 39 more

License

MIT