# superlinked/sie architecture

> Architecture diagram of the GitHub repository superlinked/sie, generated by GitDiagram from its file tree, README and a sample of source files (last updated 2026-10-02). It is an AI-made overview: check details against the source.

- Interactive diagram: https://gitdiagram.com/superlinked/sie
- Repository: https://github.com/superlinked/sie

## Overview

SIE is a self-hosted inference platform: clients submit embedding, generation, extraction, and related model tasks; server runtimes execute model work, while an optional gateway manages cluster-facing APIs and worker pools. The repository also provides SDKs, MCP tools, and framework/vector-store integrations. The graph preserves documented task families and major control-plane capabilities; some internal links are dashed because the relevant handler and routing implementations were not excerpted.

Main entry points: [`packages/sie_gateway/src/server.rs`](https://github.com/superlinked/sie/blob/main/packages/sie_gateway/src/server.rs) (Gateway HTTP server), [`packages/sie_server/src/sie_server/core/worker/model_worker.py`](https://github.com/superlinked/sie/blob/main/packages/sie_server/src/sie_server/core/worker/model_worker.py) (Model workers), [`packages/sie_sdk/src/sie_sdk`](https://github.com/superlinked/sie/tree/main/packages/sie_sdk/src/sie_sdk) (Python SDK), [`packages/sie_server/src/sie_server/api`](https://github.com/superlinked/sie/tree/main/packages/sie_server/src/sie_server/api) (Inference APIs)

## Diagram (Mermaid)

```mermaid
flowchart TD

subgraph group_runtime["Inference runtime"]
  node_server_api["Inference APIs"]
  node_encode["Encoding pipeline<br/>[encode_pipeline.py]"]
  node_model_worker["Model workers<br/>[model_worker.py]"]
  node_adapters["Model adapters"]
  node_engine_config["Engine configuration<br/>[engine.py]"]
  node_sidecar["Inference sidecar<br/>[local_ingest.rs]"]
  node_sidecar_scheduler["Batch scheduler<br/>[engine.rs]"]
end

subgraph group_control["Cluster control plane"]
  node_gateway["Gateway HTTP server<br/>[server.rs]"]
  node_gateway_handlers["Gateway handlers"]
  node_pool_manager["Pool management<br/>[pool_manager.rs]"]
  node_gateway_config["Gateway configuration<br/>[config.rs]"]
  node_nats_manager["NATS config sync<br/>[manager.rs]"]
end

subgraph group_clients["Client interfaces"]
  node_python_sdk["Python SDK"]
  node_typescript_sdk["TypeScript SDK<br/>[client.ts]"]
  node_mcp["MCP tools<br/>[server.py]"]
end

subgraph group_ecosystem["Framework integrations"]
  node_integrations["Framework adapters"]
end

node_agent(("Agent or app"))
node_nats["NATS"]

node_agent -->|"submits tasks"| node_server_api
node_server_api -.->|"dispatches encoding"| node_encode
node_encode -->|"uses worker handlers"| node_model_worker
node_model_worker -->|"invokes adapters"| node_adapters
node_engine_config -.->|"configures runtime"| node_model_worker
node_sidecar -.->|"dispatches batches"| node_sidecar_scheduler
node_gateway -->|"routes requests"| node_gateway_handlers
node_gateway -.->|"manages pools"| node_pool_manager
node_gateway -.->|"loads settings"| node_gateway_config
node_nats -->|"publishes config"| node_nats_manager
node_agent -->|"calls"| node_python_sdk
node_agent -.->|"calls"| node_typescript_sdk
node_mcp -->|"uses client"| node_python_sdk
node_integrations -->|"wraps client"| node_python_sdk

click node_server_api "https://github.com/superlinked/sie/tree/main/packages/sie_server/src/sie_server/api"
click node_encode "https://github.com/superlinked/sie/blob/main/packages/sie_server/src/sie_server/core/encode_pipeline.py"
click node_model_worker "https://github.com/superlinked/sie/blob/main/packages/sie_server/src/sie_server/core/worker/model_worker.py"
click node_adapters "https://github.com/superlinked/sie/tree/main/packages/sie_server/src/sie_server/adapters"
click node_engine_config "https://github.com/superlinked/sie/blob/main/packages/sie_server/src/sie_server/config/engine.py"
click node_sidecar "https://github.com/superlinked/sie/blob/main/packages/sie_server_sidecar/src/local_ingest.rs"
click node_sidecar_scheduler "https://github.com/superlinked/sie/blob/main/packages/sie_server_sidecar/src/scheduler/engine.rs"
click node_gateway "https://github.com/superlinked/sie/blob/main/packages/sie_gateway/src/server.rs"
click node_gateway_handlers "https://github.com/superlinked/sie/tree/main/packages/sie_gateway/src/handlers"
click node_pool_manager "https://github.com/superlinked/sie/blob/main/packages/sie_gateway/src/state/pool_manager.rs"
click node_gateway_config "https://github.com/superlinked/sie/blob/main/packages/sie_gateway/src/config.rs"
click node_nats_manager "https://github.com/superlinked/sie/blob/main/packages/sie_gateway/src/nats/manager.rs"
click node_python_sdk "https://github.com/superlinked/sie/tree/main/packages/sie_sdk/src/sie_sdk"
click node_typescript_sdk "https://github.com/superlinked/sie/blob/main/packages/sie_ts_sdk/src/client.ts"
click node_mcp "https://github.com/superlinked/sie/blob/main/packages/sie_mcp/src/sie_mcp/server.py"
click node_integrations "https://github.com/superlinked/sie/tree/main/integrations"

classDef toneNeutral fill:#f8fafc,stroke:#334155,stroke-width:1.5px,color:#0f172a
classDef toneBlue fill:#dbeafe,stroke:#2563eb,stroke-width:1.5px,color:#172554
classDef toneAmber fill:#fef3c7,stroke:#d97706,stroke-width:1.5px,color:#78350f
classDef toneMint fill:#dcfce7,stroke:#16a34a,stroke-width:1.5px,color:#14532d
classDef toneRose fill:#ffe4e6,stroke:#e11d48,stroke-width:1.5px,color:#881337
classDef toneIndigo fill:#e0e7ff,stroke:#4f46e5,stroke-width:1.5px,color:#312e81
classDef toneTeal fill:#ccfbf1,stroke:#0f766e,stroke-width:1.5px,color:#134e4a
class node_server_api,node_encode,node_model_worker,node_adapters,node_engine_config,node_sidecar,node_sidecar_scheduler toneBlue
class node_gateway,node_gateway_handlers,node_pool_manager,node_gateway_config,node_nats_manager toneAmber
class node_python_sdk,node_typescript_sdk,node_mcp toneMint
class node_integrations toneRose
class node_agent,node_nats toneIndigo
```

## Components

### Inference runtime

- [Inference APIs](https://github.com/superlinked/sie/tree/main/packages/sie_server/src/sie_server/api): `packages/sie_server/src/sie_server/api`
- [Encoding pipeline](https://github.com/superlinked/sie/blob/main/packages/sie_server/src/sie_server/core/encode_pipeline.py): `packages/sie_server/src/sie_server/core/encode_pipeline.py`
- [Model workers](https://github.com/superlinked/sie/blob/main/packages/sie_server/src/sie_server/core/worker/model_worker.py): `packages/sie_server/src/sie_server/core/worker/model_worker.py`
- [Model adapters](https://github.com/superlinked/sie/tree/main/packages/sie_server/src/sie_server/adapters): `packages/sie_server/src/sie_server/adapters`
- [Engine configuration](https://github.com/superlinked/sie/blob/main/packages/sie_server/src/sie_server/config/engine.py): `packages/sie_server/src/sie_server/config/engine.py`
- [Inference sidecar](https://github.com/superlinked/sie/blob/main/packages/sie_server_sidecar/src/local_ingest.rs): `packages/sie_server_sidecar/src/local_ingest.rs`
- [Batch scheduler](https://github.com/superlinked/sie/blob/main/packages/sie_server_sidecar/src/scheduler/engine.rs): `packages/sie_server_sidecar/src/scheduler/engine.rs`

### Cluster control plane

- [Gateway HTTP server](https://github.com/superlinked/sie/blob/main/packages/sie_gateway/src/server.rs): `packages/sie_gateway/src/server.rs`
- [Gateway handlers](https://github.com/superlinked/sie/tree/main/packages/sie_gateway/src/handlers): `packages/sie_gateway/src/handlers`
- [Pool management](https://github.com/superlinked/sie/blob/main/packages/sie_gateway/src/state/pool_manager.rs): `packages/sie_gateway/src/state/pool_manager.rs`
- [Gateway configuration](https://github.com/superlinked/sie/blob/main/packages/sie_gateway/src/config.rs): `packages/sie_gateway/src/config.rs`
- [NATS config sync](https://github.com/superlinked/sie/blob/main/packages/sie_gateway/src/nats/manager.rs): `packages/sie_gateway/src/nats/manager.rs`

### Client interfaces

- [Python SDK](https://github.com/superlinked/sie/tree/main/packages/sie_sdk/src/sie_sdk): `packages/sie_sdk/src/sie_sdk`
- [TypeScript SDK](https://github.com/superlinked/sie/blob/main/packages/sie_ts_sdk/src/client.ts): `packages/sie_ts_sdk/src/client.ts`
- [MCP tools](https://github.com/superlinked/sie/blob/main/packages/sie_mcp/src/sie_mcp/server.py): `packages/sie_mcp/src/sie_mcp/server.py`

### Framework integrations

- [Framework adapters](https://github.com/superlinked/sie/tree/main/integrations): `integrations`

### Other components

- **Agent or app**
- **NATS**

## Connections

- Agent or app → Inference APIs: submits tasks (evidence: `README.md`)
- Inference APIs → Encoding pipeline: dispatches encoding
- Encoding pipeline → Model workers: uses worker handlers (evidence: `packages/sie_server/src/sie_server/core/encode_pipeline.py`)
- Model workers → Model adapters: invokes adapters (evidence: `packages/sie_server/src/sie_server/core/worker/model_worker.py`)
- Engine configuration → Model workers: configures runtime
- Inference sidecar → Batch scheduler: dispatches batches
- Gateway HTTP server → Gateway handlers: routes requests (evidence: `packages/sie_gateway/src/server.rs`)
- Gateway HTTP server → Pool management: manages pools (evidence: `packages/sie_gateway/src/server.rs`)
- Gateway HTTP server → Gateway configuration: loads settings (evidence: `packages/sie_gateway/src/server.rs`)
- NATS → NATS config sync: publishes config (evidence: `packages/sie_gateway/src/nats/manager.rs`)
- Agent or app → Python SDK: calls (evidence: `README.md`)
- Agent or app → TypeScript SDK: calls
- MCP tools → Python SDK: uses client (evidence: `packages/sie_mcp/src/sie_mcp/server.py`)
- Framework adapters → Python SDK: wraps client (evidence: `README.md`)
