> For the complete documentation index, see [llms.txt](https://docs.autonomi.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.autonomi.com/developers/architecture/system-overview.md).

# System Overview

This page gives a high-level view of how the main developer-facing and network-facing repos fit together.

## High-level architecture

```
Application
  |                         Rust / CLI
  | REST or gRPC                |
  v                             v
 antd                      ant-core / ant
  |                             |
  +-------------+---------------+
                |
                v
             ant-node
                |
                v
            saorsa-core
                |
                v
         saorsa-transport
                |
                v
            Autonomi peers
```

## Developer interfaces

The main developer entry points are:

* `ant-sdk`, centered on the `antd` daemon plus language SDKs
* `ant-client`, centered on `ant-core` and the `ant` CLI

These interfaces target the same network, but one uses a local daemon and the other connects directly from Rust or CLI code.

## Network and node layer

The `ant-node` crate builds on `saorsa-core::P2PNode`. It adds configuration, chunk storage, payment verification, replication, upgrade handling, and node runtime management on top of the core P2P layer.

`ant-core` and `ant-node` also share wire-message types, payment-proof types, and devnet manifest types through `ant-protocol`.

In the repos used for this page, `ant-node` documents chunk storage as its active network data type.

## Routing and transport

`saorsa-core` provides the P2P node, DHT, bootstrap handling, trust system, and routing-table logic around typed addresses and peer identity. It is also responsible for publishing dialable self-addresses and coordinator hints into DHT records. `saorsa-transport` provides QUIC transport, NAT traversal, address discovery, and relay fallback where direct hole punching is not enough.

The transport story includes:

* observed-address discovery and address propagation across the network
* reachability tracking by scope: loopback, local-network, or global
* best-effort UPnP port mapping as an extra public candidate on compatible routers
* peer-ID-based hole-punch coordination, with coordinator hints and coordinator rotation when one path cannot help
* relay fallback for some CGNAT cases
* QUIC-based post-quantum transport security centered on ML-KEM-768 and ML-DSA-65

That post-quantum framing applies to transport identity and session establishment. Content self-encryption remains a separate client-side layer built around BLAKE3 and ChaCha20-Poly1305.

## Data path

On the client side, `self_encryption` turns uploaded content into a `DataMap` plus encrypted chunks. Developer-facing clients such as `antd` and `ant-core` prepare and retrieve that content, while `ant-node` stores and serves chunk data across the network.

## Upstream sources

* [ant-sdk](https://github.com/WithAutonomi/ant-sdk)
* [ant-client](https://github.com/WithAutonomi/ant-client)
* [ant-node](https://github.com/WithAutonomi/ant-node)
* [ant-protocol](https://github.com/WithAutonomi/ant-protocol)
* [self\_encryption](https://github.com/WithAutonomi/self_encryption)
* [saorsa-core](https://github.com/saorsa-labs/saorsa-core)
* [saorsa-transport](https://github.com/saorsa-labs/saorsa-transport)
* [saorsa-pqc](https://github.com/saorsa-labs/saorsa-pqc)

## Related pages

* [Core Concepts Overview](/developers/core-concepts/overview.md)
* [SDK Overview](/developers/sdk/install/reference/overview.md)
* [Using the Autonomi CLI](/developers/cli/use-the-cli.md)
* [Developing in Rust](/developers/developing-in-rust/rust.md)