Crate ai_model_vault 

AI Model Vault - Universal secure vault for AI model formats

A cross-platform, XDG-compliant secure storage system for Neural and Neurosymbolic AI models with:

• FIPS 140-3 compliant encryption
• CVE scanning and compliance
• MITRE ATT&CK framework alignment
• CMMC 2.0 compliance
• Version control with complete checkpoint history
• Format conversion capabilities

Re-exports

pub use config::VaultConfig;

pub use conversion::ConversionOptions;

pub use conversion::ConversionPipeline;

pub use conversion::ConversionProgress;

pub use conversion::ConversionResult;

pub use conversion::Converter;

pub use conversion::ValidationCheck;

pub use conversion::ValidationReport;

pub use crypto::streaming::decrypt_chunked;

pub use crypto::streaming::encrypt_chunked;

pub use crypto::streaming::is_chunked_format;

pub use crypto::streaming::StreamHeader;

pub use crypto::streaming::DEFAULT_CHUNK_SIZE;

pub use crypto::streaming::HEADER_SIZE;

pub use crypto::streaming::STREAM_MAGIC;

pub use crypto::streaming::STREAM_VERSION;

pub use error::ConversionError;

pub use error::CryptoError;

pub use error::Result;

pub use error::StorageError;

pub use error::VaultError;

pub use error::EXIT_AUTH;

pub use error::EXIT_COMPLIANCE;

pub use error::EXIT_CONFIG;

pub use error::EXIT_GENERAL;

pub use error::EXIT_INTEGRITY;

pub use error::EXIT_INVALID_INPUT;

pub use error::EXIT_NOT_FOUND;

pub use error::EXIT_PERMISSION;

pub use error::EXIT_SUCCESS;

pub use model_card::CaveatsAndRecommendations;

pub use model_card::EnvironmentalImpact;

pub use model_card::EthicalConsiderations;

pub use model_card::Evaluation;

pub use model_card::IntendedUse;

pub use model_card::Metric;

pub use model_card::ModelCard;

pub use model_card::ModelDetails;

pub use model_card::TrainingData;

pub use rag::Database;

pub use rag::Document;

pub use rag::DocumentStore;

pub use rag::InMemoryDatabase;

pub use rag::KnowledgeBase;

pub use rag::KnowledgeBaseConfig;

pub use rag::MCPServer;

pub use rag::MCPTool;

pub use rag::RetrievalCache;

pub use rag::Rule;

pub use rag::RuleAction;

pub use rag::RuleCondition;

pub use rag::RuleEngine;

pub use rag::ToolContext;

pub use rag::ToolExecutor;

pub use rag::ToolResult;

pub use traits::AimvUri;

pub use traits::AsyncBlobStore;

pub use traits::AsyncBlobStoreAdapter;

pub use traits::AuditLogSubscriber;

pub use traits::AuditSink;

pub use traits::BlobInfo;

pub use traits::BlobReceipt;

pub use traits::BlobStore;

pub use traits::BlobStoreStats;

pub use traits::CryptoProvider;

pub use traits::EventBus;

pub use traits::EventSubscriber;

pub use traits::MetricsSnapshot;

pub use traits::MetricsSubscriber;

pub use traits::NullAuditSink;

pub use traits::VaultEvent;

pub use traits::VaultMetrics;

pub use traits::VaultState;

pub use traits::VersionRepo;

pub use utils::CompressionAnalyzer;

pub use utils::CompressionReport;

pub use utils::ModelAnalysis;

pub use utils::ModelAnalyzer;

pub use utils::ModelArchive;

pub use utils::ModelDeduplicator;

pub use utils::ModelExporter;

pub use utils::PruningInfo;

pub use utils::PruningMethod;

pub use utils::QuantizationInfo;

pub use utils::QuantizationSavings;

pub use utils::RetrievalOptimizer;

pub use vault::Vault;

pub use vault::VaultBuilder;

pub use vault::VersionBackend;

pub use version::ModelVersion;

pub use version::VersionControl;

pub use version_sqlite::SqliteVersionRepo;

pub use benchmark::BenchmarkRecord;

pub use benchmark::BenchmarkResult;

pub use benchmark::BenchmarkStore;

pub use diff::DiffSummary;

pub use diff::ModelDiff;

pub use diff::ModelDiffer;

pub use diff::TensorChange;

pub use diff::TensorInfo;

pub use download::ModelDownloader;

pub use download::ModelSource;

pub use interop::register_lm_studio;

pub use interop::register_ollama;

pub use interop::InferenceEngine;

pub use interop::LmStudioOptions;

pub use interop::OllamaOptions;

pub use interop::RegistrationResult;

pub use license_scan::DetectedLicense;

pub use license_scan::LicenseClass;

pub use license_scan::LicenseScanReport;

pub use license_scan::LicenseScanner;

pub use license_scan::LicenseSource;

pub use scanning::PickleScanner;

pub use scanning::ScanFinding;

pub use scanning::ScanReport;

pub use scanning::Severity;

pub use signing::ModelSignature;

pub use signing::ModelSigner;

pub use signing::SignatureVerification;

pub use signing::SigningKeyPair;

pub use blockchain::AuditBlock;

pub use blockchain::AuditProof;

pub use blockchain::BlockchainAudit;

pub use blockchain::ChainVerification;

pub use blockchain::MerkleProof;

pub use blockchain::MerkleTree;

pub use blockchain::ProofVerification;

pub use federation::ConflictResolution;

pub use federation::FederationConfig;

pub use federation::FederationManager;

pub use federation::FederationStatus;

pub use federation::PeerConfig;

pub use federation::SyncConflict;

pub use federation::SyncManifest;

pub use federation::SyncResult;

pub use federation::VectorClock;

pub use access_control::AclEntry;

pub use access_control::AclGuard;

pub use access_control::Role;

pub use gc::GcReport;

pub use kms::is_kms_uri;

pub use kms::KmsBackend;

pub use kms::KmsRequest;

pub use kms::KmsUri;

pub use lineage_graph::DerivationKind;

pub use lineage_graph::LineageEdge;

pub use lineage_graph::LineageGraph;

pub use plugins::PluginInfo;

pub use plugins::PluginManifest;

pub use plugins::PluginRegistry;

pub use policies::PolicyReport;

pub use policies::PolicyStore;

pub use policies::RetentionPolicy;

pub use profiles::Profile;

pub use profiles::ProfileStore;

pub use tags::SearchQuery;

pub use tags::SearchResult;

pub use tags::TagStore;

pub use validation::ValidationProbe;

pub use validation::ValidationResult;

pub use validation::ValidationStore;

pub use vault_bundle::BundleManifest;

pub use vault_bundle::ExportReport;

pub use vault_bundle::ImportReport;

pub use webhooks::WebhookPayload;

pub use webhooks::WebhookStore;

pub use webhooks::WebhookTarget;

pub use quantization::BatchQuantReport;

pub use quantization::QuantMethod;

pub use quantization::QuantProfile;

pub use quantization::QuantProfileStore;

pub use quantization::QuantResult;

pub use evaluation::EvalComparison;

pub use evaluation::EvalRun;

pub use evaluation::EvalStore;

pub use evaluation::MetricDelta;

pub use evaluation::MetricResult;

pub use scheduler::BackupFrequency;

pub use scheduler::BackupManager;

pub use scheduler::BackupRecord;

pub use scheduler::BackupReport;

pub use scheduler::BackupSchedule;

pub use multi_vault::VaultEntry;

pub use multi_vault::VaultRegistry;

pub use multi_vault::VaultSummary;

Modules

access_control

Access control — role-based permissions for multi-user vaults.

api

REST API server for AI Model Vault

audit

Audit logging for compliance and security monitoring

benchmark

Benchmark metadata — attach evaluation results to model versions.

blockchain

Blockchain-based immutable audit trail

cloud_envelope

Client-side encryption for objects leaving the machine.

compliance

Compliance checking and security validation

config

XDG Base Directory Specification compliant configuration

conversion

Model format conversion pipeline

crypto

FIPS 140-3 Compliant Cryptographic Module

diff

Model diffing — compare two versions of a model.

download

Model download from trusted sources.

error

Error types for AI Model Vault

evaluation

Model evaluation harness.

federation

Federated vault synchronization

federation_transport

Shared plumbing for the federation client and server halves.

formats

Model format detection and conversion registry

gc

Garbage collection — find and remove orphaned blobs, temp files, and unreferenced data inside a vault directory.

gguf

Minimal GGUF header reader.

interop

Inference engine interop — register vault models with Ollama and LM Studio.

kms

Secrets manager integration — derive vault passphrases from external KMS.

license_scan

License scanning — detect model licenses from metadata and model cards.

lineage_graph

Cross-model lineage graph — track derivation relationships between different models (fine-tunes, merges, distillations, quantisations).

model_card

Model Card implementation following industry standards

multi_vault

Multi-vault management.

permissions

Cross-platform file permission utilities.

plugins

Plugin system — dynamic extension loading for custom commands and hooks.

policies

Snapshot / retention policies — automated version cleanup rules.

profiles

Configuration profiles — named sets of config overrides.

quantization

Model quantization pipeline.

rag

RAG (Retrieval-Augmented Generation) and Rule-Based System Support

scanning

Pickle safety scanner for PyTorch model files.

scheduler

Vault backup scheduling.

signing

Ed25519 model signing and provenance verification.

storage

Storage backend for encrypted model data

tags

Model tagging and search — attach tags and query models by metadata.

telemetry

Telemetry and analytics module for AI Model Vault.

telemetry_otlp

OTLP export transport for telemetry events. OTLP export for telemetry events (otel feature).

traits

Core trait definitions for AI Model Vault architecture v2.

tui

Terminal UI — interactive model browser for the vault.

utils

AI Model Utilities - Common operations for AI models

validation

Model inference validation — sanity-check model outputs after storage or conversion to catch silent corruption.

vault

Main Vault implementation

vault_bundle

Vault export/import — portable encrypted vault bundles.

version

Version control system for model checkpoints

version_sqlite

SQLite-backed version repository.

webhooks

Webhook / notification system — fire HTTP callbacks on vault events.

Constants

VERSION

Library version