How alvaritoo89/locall works
Locall is a Spanish-language, local-first AI model manager and inference platform. Users discover compatible models, install or import them, then chat through the CLI, desktop app, or local API; models and conversations remain on the machine. The API mediates client requests to the Rust core, which coordinates model storage, registry, scheduling, and inference. The README also describes signed model catalogs and a website. Some subsystem links below are based on README-level architecture because their implementation excerpts were not sampled.
Main entry points: src/lib.rs Core service · src/lib.rs Local API · src/lib.rs Model scheduler · main.rs Desktop bridge
Components 19
User clients
- Desktop bridge
crates/locall-app/src/main.rs - Desktop UI
crates/locall-app/ui/src/App.tsx - CLI
crates/locall-cli/src/ordenes.rs
Local service
- Local API
crates/locall-api/src/lib.rs - Core service
crates/locall-core/src/lib.rs - Settings and logs
crates/locall-config/src/lib.rs
Model lifecycle
- Signed catalog
crates/locall-catalogo/src/lib.rs - Hugging Face discovery
crates/locall-core/src/huggingface.rs - Compatibility checks
crates/locall-compat/src/lib.rs - GGUF reader
crates/locall-gguf/src/lib.rs - Verified downloads
crates/locall-download/src/lib.rs - Model store
crates/locall-store/src/lib.rs - Model and chat registry
crates/locall-registry/src/lib.rs
Inference runtime
- Model scheduler
crates/locall-scheduler/src/lib.rs - Inference engine control
crates/locall-engine/src/lib.rs - Machine and process support
crates/locall-system/src/lib.rs
Other components
- User
- llama.cpp process
- Project website
crates/locall-web/src/lib.rs
Connections 20
- User → Desktop bridge: uses
- User → CLI: uses
- User → Local API: uses
- Desktop UI → Desktop bridge: invokes
crates/locall-app/src/main.rs - Desktop bridge → Local API: requests
crates/locall-app/src/main.rs - CLI → Local API: requests
- Local API → Core service: delegates to
crates/locall-api/src/lib.rs - Core service → Settings and logs: opens
crates/locall-core/src/lib.rs - Core service → Signed catalog: uses
- Core service → Hugging Face discovery: discovers models
crates/locall-core/src/lib.rs - Core service → Compatibility checks: checks fit
crates/locall-core/src/lib.rs - Core service → GGUF reader: reads headers
crates/locall-core/src/lib.rs - Core service → Verified downloads: downloads models
- Core service → Model store: manages files
- Core service → Model and chat registry: records models and chats
crates/locall-core/src/lib.rs - Core service → Model scheduler: requests model use
- Model scheduler → Inference engine control: starts and stops
crates/locall-scheduler/src/lib.rs - Inference engine control → llama.cpp process: controls process
- Model scheduler → Machine and process support: checks memory
crates/locall-scheduler/src/lib.rs - User → Project website: visits