# Store and Forward

## What Is Store and Forward?

Store and Forward (S&amp;F) is a Meshtastic server module that **buffers text messages for nodes that are temporarily offline**. When a client node comes back within range, the S&amp;F server can replay text messages the client missed while it was away.

**Important - the buffer is volatile and S&amp;F is not a delivery guarantee.** Stored messages live only in the server node's PSRAM. Any reboot, brownout, or power loss on the server node erases ALL stored messages with no recovery. Replay is best-effort, request-driven, and limited by count and time window; it covers text messages only (not telemetry, position, or other packet types). Never treat the buffer as durable storage, and never rely on it for life-safety message assurance - for incident use, assume the buffer can vanish at any moment.

## How It Works

1. A node with S&amp;F enabled in *server mode* listens to text messages on its primary channel (channel 0) and stores them in a ring buffer in PSRAM. Note: history retrieval over LoRa is **not available on the default public channel**, so the server must be configured with a non-default primary channel for S&amp;F history to work.
2. When a client node requests history (see Client Configuration below), it sends a replay request.
3. The server transmits buffered messages up to the configured **History Return Max** count and within the **History Return Window** time range.

## Hardware Requirements

The S&amp;F server requires an ESP32 device **with PSRAM** - nRF52-based boards cannot run the Store-and-Forward server module at all:

- **Recommended (ESP32-with-PSRAM only):** T-Beam (v1.0 and later), LilyGo T3S3, or other ESP32-S3 devices that include PSRAM.
- **Not supported as an S&amp;F server:** nRF52840-based devices (RAK4631, Nordic development kits) and any board without PSRAM. These cannot act as a Store-and-Forward server.

## Server Configuration

1. Navigate to **Config → Module Config → Store and Forward**.
2. Set **Enabled → true**.
3. Set **Is Server → true**.
4. Set **History Return Max** - the maximum number of messages to replay per request (default: 25 messages, which is the value used when this is left at 0).
5. Set **History Return Window** - the time window in minutes from which to replay messages (default 0 = 240 minutes / 4 hours).
6. Save and reboot the node.

## Client Configuration

A client node does **not** need the Store and Forward module enabled to request history. To retrieve missed messages:

1. Set role to **CLIENT** (or **CLIENT\_MUTE**, labelled "Client Mute" in the app menu, if the device should not rebroadcast).
2. Request history manually: on Android, send a direct message containing **SF** to the server node; on Apple, use the **Client History** option.

History retrieval over LoRa is manually requested as above. Automatic retrieval happens only when an app connects directly to the server node (firmware 2.4+) - it does not happen automatically over the mesh just by coming into range.

## Ideal Use Cases

- Hikers or cyclists who move in and out of mesh coverage.
- Vehicles that periodically leave and re-enter a covered area.
- Remote monitoring stations that reconnect on a schedule.
- Community nets that want a convenience replay of recent text traffic - but note S&amp;F is best-effort only and must never be relied on as guaranteed delivery for emergency or life-safety messages.

## Limitations

- S&amp;F history retrieval over LoRa is **not available on the default public channel**; you must configure a non-default primary channel to use it. S&amp;F operates only on the primary channel (channel 0) - if your net runs on a private or secondary channel, store-and-forward will not buffer or replay that traffic. Verify replay actually works on YOUR channel before depending on it.
- Text messages only - telemetry, position, and other packet types are not buffered or replayed.
- Large message volumes will cause the ring buffer to overwrite older messages quickly.
- **The buffer lives in PSRAM only.** The server node must be continuously powered and on-mesh; any reboot, brownout, or power loss erases ALL stored messages with no recovery.
- S&amp;F adds overhead to the server node; monitor channel utilisation on high-traffic networks.