diff --git a/.github/ISSUE_TEMPLATE/bug_report.yml b/.github/ISSUE_TEMPLATE/bug_report.yml
index c10026e..5f67fea 100644
--- a/.github/ISSUE_TEMPLATE/bug_report.yml
+++ b/.github/ISSUE_TEMPLATE/bug_report.yml
@@ -49,10 +49,10 @@ body:
       label: "Environment"
       description: "Your setup details."
       placeholder: |
-        - OS: macOS 15 / Ubuntu 24.04 / Raspberry Pi OS
+        - OS: macOS 15 / Ubuntu 24.04 / Debian 12 / Raspberry Pi OS
         - Python version: 3.11.x
         - Node version: 20.x
-        - Deployment mode: local dev / Pi production / other
+        - Deployment mode: local dev / Linux-host production / other
         - VocalizeAI version / commit hash:
     validations:
       required: true
@@ -61,7 +61,7 @@ body:
     id: logs
     attributes:
       label: "Relevant logs"
-      description: "Paste any relevant log output. Use `journalctl -u vocalize` on Pi, or the uvicorn terminal output locally."
+      description: "Paste any relevant log output. Use `journalctl -u vocalize` on a systemd-installed host, or the uvicorn terminal output locally."
       render: shell
     validations:
       required: false
diff --git a/CONTRIBUTING.md b/CONTRIBUTING.md
index 4341644..b8b30e5 100644
--- a/CONTRIBUTING.md
+++ b/CONTRIBUTING.md
@@ -37,7 +37,7 @@ cd frontend && npm run test:integration
 ```
 
 Note: `tests/integration/` release-audio cases require a physical audio setup
-(microphone + speaker) and a live Pi orchestrator. These are gated behind
+(microphone + speaker) and a live Linux-host orchestrator. These are gated behind
 `--release-audio` and do not run in PR CI.
 
 All checks must pass on your PR before merge. CI runs lint (ruff + mypy + tsc),
diff --git a/README.md b/README.md
index ed2aec1..a74d0da 100644
--- a/README.md
+++ b/README.md
@@ -10,8 +10,9 @@ across languages when needed.
 
 ## Current Status
 
-**v1 ships** the universal phone-task engine, Web console, and Raspberry Pi
-orchestrator deployment. The backend 5-layer prompt architecture
+**v1 ships** the universal phone-task engine, Web console, and a Linux-host
+orchestrator deployment (Raspberry Pi was the original reference target;
+any modern Linux host with systemd works). The backend 5-layer prompt architecture
 (task_planner / preflight / merchant_agent / clarification_collector / relay)
 handles any phone task — restaurant bookings, service appointments, balance
 inquiries, status checks, and more. An OSS mirror is available at
@@ -91,12 +92,12 @@ VocalizeAI/
 │   ├── messages/              # next-intl zh/en bundles
 │   └── tests/                 # vitest unit tests
 ├── demos/                     # runnable demos
-├── infra/                     # deployment scripts (GPU node, Pi orchestrator)
+├── infra/                     # deployment scripts (GPU node, Linux orchestrator)
 ├── tests/                     # pytest suite
 │   └── integration/           # Playwright laptop-loopback + AI-merchant harness
 ├── install/                   # one-shot install scripts
 │   ├── dev-install.sh         # Mac/Linux local dev setup
-│   └── pi-install.sh          # Raspberry Pi production deploy
+│   └── install.sh             # Linux production deploy (Raspberry Pi is one example target)
 ├── docs/                      # architecture, deploy guides, release evidence
 ├── scripts/                   # smoke test and utility scripts
 │   └── smoke.sh               # post-install end-to-end verification
@@ -117,13 +118,14 @@ VocalizeAI/
 See `.env.example` for the full env-var inventory including LLM, GPU service,
 and frontend build-time variables.
 
-For the full production Pi deployment runbook, see [docs/deploy/pi.md](docs/deploy/pi.md).
+For the full Linux-host production deployment runbook (Raspberry Pi is one
+example target), see [docs/deploy/linux.md](docs/deploy/linux.md).
 
 ### GPU node requirements
 
 SenseVoice (STT) and CosyVoice (TTS) run as separate GPU services and connect
-to the Pi orchestrator over Tailscale. GPU services are optional for local dev
-(the LLM path works without them). See [docs/deploy/pi.md](docs/deploy/pi.md)
+to the orchestrator host over Tailscale. GPU services are optional for local dev
+(the LLM path works without them). See [docs/deploy/linux.md](docs/deploy/linux.md)
 for the GPU node setup.
 
 ## Run the dev server
diff --git a/README.zh-CN.md b/README.zh-CN.md
index 570dcf1..2382df1 100644
--- a/README.zh-CN.md
+++ b/README.zh-CN.md
@@ -85,12 +85,12 @@ VocalizeAI/
 │   ├── messages/              # next-intl zh/en bundles
 │   └── tests/                 # vitest unit tests
 ├── demos/                     # runnable demos
-├── infra/                     # deployment scripts (GPU node, Pi orchestrator)
+├── infra/                     # 部署脚本(GPU 节点、Linux 编排器)
 ├── tests/                     # pytest suite
 │   └── integration/           # Playwright laptop-loopback + AI-merchant harness
 ├── install/                   # 一键安装脚本
 │   ├── dev-install.sh         # Mac/Linux 本地开发环境安装
-│   └── pi-install.sh          # 树莓派生产部署安装
+│   └── install.sh             # Linux 生产部署安装(树莓派是一种受支持的目标)
 ├── docs/                      # 架构文档、部署指南、发布记录
 ├── scripts/                   # smoke 测试和工具脚本
 │   └── smoke.sh               # 安装后端到端验证脚本
@@ -110,13 +110,13 @@ VocalizeAI/
 
 完整环境变量清单（含 LLM、GPU 服务、前端构建变量）见 `.env.example`。
 
-完整的树莓派生产部署手册，见 [docs/deploy/pi.md](docs/deploy/pi.md)。
+完整的树莓派生产部署手册，见 [docs/deploy/linux.md](docs/deploy/linux.md)。
 
 ### GPU 节点要求
 
 SenseVoice（STT）和 CosyVoice（TTS）作为独立 GPU 服务运行，通过 Tailscale
 与树莓派编排器连接。本地开发不需要 GPU（只需 LLM 路径即可运行）。GPU 节点配置见
-[docs/deploy/pi.md](docs/deploy/pi.md)。
+[docs/deploy/linux.md](docs/deploy/linux.md)。
 
 ## 跑开发服务器
 
diff --git a/docs/architecture.md b/docs/architecture.md
index b080669..01f130c 100644
--- a/docs/architecture.md
+++ b/docs/architecture.md
@@ -556,7 +556,7 @@ See: `src/vocalize/server/ws.py`, `frontend/lib/audio*`, `frontend/components/Br
 | Env/config loading | `src/vocalize/config.py` |
 | Asyncio main pipeline | `src/vocalize/pipeline.py` |
 | Frontend (Next.js 14) | `frontend/` |
-| Pi deployment assets | `infra/pi-orchestrator/` |
+| Pi deployment assets | `infra/orchestrator/` |
 | GPU services setup | `infra/gpu-services/` |
 | Backend tests (pytest) | `tests/` |
 | Integration tests (Playwright) | `tests/integration/` |
@@ -611,5 +611,5 @@ See: `src/vocalize/server/ws.py`, `src/vocalize/server/sessions.py`
 ## Further Reading
 
 - **[docs/deploy/local.md](docs/deploy/local.md)** — Mac/Linux dev environment setup and env-var reference
-- **[docs/deploy/pi.md](docs/deploy/pi.md)** — End-to-end Pi production deployment runbook
+- **[docs/deploy/linux.md](docs/deploy/linux.md)** — End-to-end Pi production deployment runbook
 - **[CONTRIBUTING.md](../CONTRIBUTING.md)** — Contributor flow, code style, commit conventions
diff --git a/docs/deploy/pi.md b/docs/deploy/linux.md
similarity index 72%
rename from docs/deploy/pi.md
rename to docs/deploy/linux.md
index 1a56e20..a44a91a 100644
--- a/docs/deploy/pi.md
+++ b/docs/deploy/linux.md
@@ -1,57 +1,62 @@
-# Deploying VocalizeAI on a Raspberry Pi
+# Deploying VocalizeAI on a Linux Host
 
-This runbook covers end-to-end production deployment of VocalizeAI on a
-Raspberry Pi: the orchestrator runs on the Pi; GPU services (SenseVoice STT +
-CosyVoice TTS) run on a separate machine reachable over Tailscale; a Cloudflare
-Tunnel fronts the Pi to the public internet.
+This runbook covers end-to-end production deployment of VocalizeAI on any
+modern Linux host with systemd: the orchestrator runs on this host, GPU
+services (SenseVoice STT + CosyVoice TTS) run on a separate machine reachable
+over Tailscale, and a Cloudflare Tunnel fronts the orchestrator host to the
+public internet.
+
+Tested on **Debian 12**, **Ubuntu 22.04 / 24.04**, and **Raspberry Pi OS
+(Bookworm)**. A Raspberry Pi was the original reference target — see
+["Hardware example: Raspberry Pi"](#hardware-example-raspberry-pi) below for
+the BOM, OS imaging, and SSH bootstrap steps for that specific target.
 
 ---
 
-## Hardware Bill of Materials
+## Bill of Materials
 
-**Raspberry Pi:**
-- Raspberry Pi 4 or Pi 5, **8 GB RAM recommended** (4 GB works for the orchestrator
-  alone but is tight if other services run alongside)
-- 32 GB+ microSD card or USB SSD (SSD strongly recommended for production)
-- Reliable internet connection (Cloudflare Tunnel requires outbound HTTPS)
+**Orchestrator host:**
+- Any 64-bit Linux box with systemd, ≥ 2 GB RAM, ≥ 16 GB free disk.
+- Python 3.11 (installed by step 1 of `install/install.sh`).
+- Persistent internet connection (Cloudflare Tunnel requires outbound HTTPS).
 
 **GPU node (separate machine):**
-- NVIDIA RTX-class GPU (GTX 1080 or better; RTX 30/40 series recommended)
-- Windows + WSL2 or Linux (PyTorch 2.7.1+cu128)
-- Reachable from the Pi over Tailscale on the configured `GPU_HOST` IP/hostname
+- NVIDIA RTX-class GPU (GTX 1080 or better; RTX 30/40 series recommended).
+- Windows + WSL2 or Linux (PyTorch 2.7.1+cu128).
+- Reachable from the orchestrator host over Tailscale on the configured
+  `GPU_HOST` IP/hostname.
 
 **Network:**
-- Tailscale account (free tier is sufficient) with both the Pi and GPU node enrolled
-- Cloudflare account with a domain pointed at Cloudflare DNS (free tier is sufficient)
+- Tailscale account (free tier is sufficient) with both the orchestrator host
+  and the GPU node enrolled.
+- Cloudflare account with a domain pointed at Cloudflare DNS (free tier is
+  sufficient).
 
 ---
 
 ## OS Preparation
 
 ```bash
-# Flash Raspberry Pi OS Lite (64-bit) to the SD card / SSD using Raspberry Pi Imager.
-# In Imager, pre-configure:
-#   - hostname
-#   - SSH enabled
-#   - SSH public key (paste your ~/.ssh/id_ed25519.pub or generate one first)
-#   - Wi-Fi credentials (if not using Ethernet)
-
-# After first boot, SSH in and update the system:
-ssh pi@<pi-hostname>
+# On the orchestrator host (any modern 64-bit Linux with systemd):
+ssh <user>@<host>
 sudo apt-get update && sudo apt-get upgrade -y
 
 # Ensure git and curl are present:
 sudo apt-get install -y git curl
 ```
 
+For the Raspberry Pi-specific imaging / first-boot steps, see
+["Hardware example: Raspberry Pi"](#hardware-example-raspberry-pi).
+
 ---
 
 ## Tailscale Setup
 
-Tailscale provides the encrypted overlay network between the Pi and the GPU node.
+Tailscale provides the encrypted overlay network between the orchestrator
+host and the GPU node.
 
 ```bash
-# Install Tailscale on the Pi:
+# Install Tailscale on the orchestrator host:
 curl -fsSL https://tailscale.com/install.sh | sh
 sudo tailscale up
 
@@ -66,7 +71,7 @@ tailscale status
 
 Set `GPU_HOST` in `/opt/vocalize/.env` to the GPU node's Tailscale IP.
 
-If the GPU services are not yet running, use `install/pi-install.sh --skip-gpu`
+If the GPU services are not yet running, use `install/install.sh --skip-gpu`
 to proceed with installation without the GPU-reachability check.
 
 ---
@@ -74,20 +79,20 @@ to proceed with installation without the GPU-reachability check.
 ## Clone and Install
 
 ```bash
-# Clone the repository to the Pi:
+# Clone the repository to the orchestrator host:
 git clone https://github.com/DGPisces/VocalizeAI.git /opt/vocalize
 cd /opt/vocalize
 
 # Dry-run first to preview all 7 steps:
-bash install/pi-install.sh --dry-run
+bash install/install.sh --dry-run
 
 # Run the full installer:
-bash install/pi-install.sh
+bash install/install.sh
 
 # Or run selectively:
-bash install/pi-install.sh --steps "1,2,6"   # apt + venv + systemd only
-bash install/pi-install.sh --skip-gpu         # skip GPU-reachability check in step 7
-bash install/pi-install.sh --skip-tunnel      # skip step 5 (Cloudflare Tunnel info)
+bash install/install.sh --steps "1,2,6"   # apt + venv + systemd only
+bash install/install.sh --skip-gpu         # skip GPU-reachability check in step 7
+bash install/install.sh --skip-tunnel      # skip step 5 (Cloudflare Tunnel info)
 ```
 
 **Installer steps:**
@@ -96,7 +101,7 @@ bash install/pi-install.sh --skip-tunnel      # skip step 5 (Cloudflare Tunnel i
 |------|--------|
 | 1 | `apt-get install` python3.11 python3.11-venv python3-pip build-essential rsync |
 | 2 | Create `.venv` in `/opt/vocalize`, `pip install -e .` |
-| 3 | GPU services note (GPU lives on a separate host; no on-Pi install) |
+| 3 | GPU services note (GPU lives on a separate host; no on-orchestrator install) |
 | 4 | Tailscale presence check (warns if absent) |
 | 5 | Cloudflare Tunnel token-install instructions |
 | 6 | Copy `vocalize.service` to `/etc/systemd/system/`, copy `.env.template` to `/opt/vocalize/.env` if absent, `systemctl enable vocalize` |
@@ -124,9 +129,9 @@ sudo nano /opt/vocalize/.env
 | `GPU_HOST` | yes (if using GPU) | STT/TTS host — Tailscale IP of your GPU node |
 | `SENSEVOICE_WS_PORT` | default ok | STT port; default `8000` |
 | `COSYVOICE_WS_PORT` | default ok | TTS port; default `8001` |
-| `VOCALIZE_HOST` | default ok | uvicorn bind host; set to `0.0.0.0` for Pi production |
+| `VOCALIZE_HOST` | default ok | uvicorn bind host; set to `0.0.0.0` for production |
 | `VOCALIZE_PORT` | default ok | uvicorn bind port; default `8080` |
-| `ORCHESTRATOR_LISTEN_PORT` | default ok | Pi service port; default `8080` (legacy compatibility) |
+| `ORCHESTRATOR_LISTEN_PORT` | default ok | Orchestrator service port; default `8080` (legacy compatibility) |
 | `VOCALIZE_WS_BASE_URL` | **yes** | Public WS base URL; e.g. `wss://api.<your-domain>` — startup raises if missing in non-localhost mode |
 | `VOCALIZE_CORS_ORIGINS` | default ok | Comma-separated allowed CORS origins; default auto-picked from VOCALIZE_HOST |
 | `DEFAULT_LANGUAGE` | default ok | `zh` or `en`; default `zh` |
@@ -153,7 +158,7 @@ VOCALIZE_CORS_ORIGINS=https://<your-domain>
 
 ## Cloudflare Tunnel
 
-The Cloudflare Tunnel connects the Pi to the public internet without exposing SSH
+The Cloudflare Tunnel connects the orchestrator host to the public internet without exposing SSH
 or opening firewall ports.
 
 **Token-based install (recommended):**
@@ -163,7 +168,7 @@ or opening firewall ports.
 # Zero Trust -> Networks -> Tunnels -> [your tunnel] -> Configure
 # -> "Install and run a connector" -> Copy the displayed token
 
-# Install the tunnel service on the Pi:
+# Install the tunnel service on the orchestrator host:
 sudo cloudflared service install <TUNNEL_TOKEN>
 
 # Verify the service is running:
@@ -171,7 +176,7 @@ sudo systemctl status cloudflared
 ```
 
 The reference ingress shape for this project is documented in
-`infra/pi-orchestrator/cloudflared-config.yml` (maps
+`infra/orchestrator/cloudflared-config.yml` (maps
 `vocalize-api.<your-tunnel-name>` → `http://localhost:8080` and
 `vocalize.<your-tunnel-name>` → `http://localhost:3000`). Configure
 the actual public hostname routing in the Cloudflare dashboard under
@@ -186,7 +191,7 @@ tunnel name; tunnels are account-specific.
 
 ### vocalize.service
 
-The `vocalize.service` unit file is at `infra/pi-orchestrator/vocalize.service`
+The `vocalize.service` unit file is at `infra/orchestrator/vocalize.service`
 and is copied to `/etc/systemd/system/vocalize.service` by step 6 of the installer.
 
 ```ini
@@ -240,7 +245,7 @@ sudo systemctl restart vocalize
 After the installer completes (step 7 runs this automatically), verify the deployment:
 
 ```bash
-# Smoke test against the Pi's local port:
+# Smoke test against the orchestrator's local port:
 VOCALIZE_API_BASE=http://127.0.0.1:8080 bash scripts/smoke.sh
 # Exit 0 = working deployment
 
@@ -251,7 +256,7 @@ VOCALIZE_API_BASE=https://api.<your-domain> bash scripts/smoke.sh
 The smoke script exercises 6 round-trips: `GET /health`, `POST /api/sessions`,
 `POST /api/sessions/{id}/task`, WS upgrade + send/recv, `DELETE /api/sessions/{id}`.
 
-Note: the local smoke on the Pi uses port 8080 (production port), not 8000 (dev
+Note: the local smoke uses port 8080 (production port), not 8000 (dev
 port). Make sure `VOCALIZE_API_BASE` is set accordingly.
 
 ---
@@ -311,3 +316,39 @@ Common causes:
 **Port conflicts:**
 - `VOCALIZE_PORT` defaults to 8080. If another service occupies that port, change
   `VOCALIZE_PORT` in `.env` and update the Cloudflare Tunnel ingress rule accordingly.
+
+---
+
+## Hardware example: Raspberry Pi
+
+The Raspberry Pi was the original reference target for this runbook. None of
+the steps above are Pi-specific; this section just captures the bits that
+differ when the orchestrator host happens to be a Pi.
+
+### BOM
+
+- Raspberry Pi 4 or Pi 5, **8 GB RAM recommended** (4 GB works for the
+  orchestrator alone but is tight if other services run alongside).
+- 32 GB+ microSD card or USB SSD (SSD strongly recommended for production).
+- Reliable internet connection (Cloudflare Tunnel requires outbound HTTPS).
+
+### Imaging and first boot
+
+```bash
+# Flash Raspberry Pi OS Lite (64-bit) to the SD card / SSD using Raspberry
+# Pi Imager. In Imager, pre-configure:
+#   - hostname
+#   - SSH enabled
+#   - SSH public key (paste your ~/.ssh/id_ed25519.pub or generate one first)
+#   - Wi-Fi credentials (if not using Ethernet)
+
+# After first boot, SSH in and update the system:
+ssh pi@<pi-hostname>
+sudo apt-get update && sudo apt-get upgrade -y
+
+# Ensure git and curl are present:
+sudo apt-get install -y git curl
+```
+
+From here on, the rest of this runbook (Tailscale, install, Cloudflare
+Tunnel, smoke) applies unchanged.
diff --git a/docs/deploy/local.md b/docs/deploy/local.md
index 631d5fb..b7b83b0 100644
--- a/docs/deploy/local.md
+++ b/docs/deploy/local.md
@@ -72,12 +72,12 @@ $EDITOR .env
 | `OPENAI_API_KEY` | **yes** | LLM authentication — any OpenAI-compatible provider (OpenAI, DeepSeek, Qwen, etc.) |
 | `OPENAI_BASE_URL` | default ok | LLM endpoint; default `https://api.deepseek.com/v1` |
 | `OPENAI_MODEL` | default ok | Model name; default `deepseek-chat` |
-| `GPU_HOST` | only if using GPU | STT/TTS host; use `localhost` for single-machine dev, Tailscale IP for Pi deployment |
+| `GPU_HOST` | only if using GPU | STT/TTS host; use `localhost` for single-machine dev, Tailscale IP for remote-GPU deployment (e.g. Raspberry Pi orchestrator → GPU node) |
 | `SENSEVOICE_WS_PORT` | default ok | SenseVoice STT WebSocket port; default `8000` |
 | `COSYVOICE_WS_PORT` | default ok | CosyVoice TTS WebSocket port; default `8001` |
 | `VOCALIZE_HOST` | default ok | uvicorn bind host; `127.0.0.1` for local dev, `0.0.0.0` for production |
 | `VOCALIZE_PORT` | default ok | uvicorn bind port; default `8080` (note: dev `main.py` defaults to 8000) |
-| `ORCHESTRATOR_LISTEN_PORT` | default ok | Pi service port; default `8080` (legacy; mirrors `VOCALIZE_PORT`) |
+| `ORCHESTRATOR_LISTEN_PORT` | default ok | Orchestrator service port; default `8080` (legacy; mirrors `VOCALIZE_PORT`) |
 | `VOCALIZE_WS_BASE_URL` | required when non-localhost | Public WS base URL (e.g. `wss://api.example.com`); startup raises if missing in non-localhost mode (D-11) |
 | `VOCALIZE_CORS_ORIGINS` | default ok | Comma-separated allowed CORS origins; auto-picked from VOCALIZE_HOST in dev mode |
 | `DEFAULT_LANGUAGE` | default ok | Session default language; `zh` or `en`; default `zh` |
@@ -178,7 +178,7 @@ cd frontend && npm run test:integration
 ```
 
 Note: `tests/integration/` release-audio cases require a physical audio setup
-(microphone + speaker) and a live Pi orchestrator. These are gated behind
+(microphone + speaker) and a live Linux-host orchestrator. These are gated behind
 `--release-audio` and do not run in PR CI. The standard integration test suite
 (`npm run test:integration`) runs the 8 text-bypass AI-merchant scenarios and
 does not require physical hardware.
diff --git a/infra/README.md b/infra/README.md
index 32b7f0c..3b9548c 100644
--- a/infra/README.md
+++ b/infra/README.md
@@ -5,8 +5,10 @@ home of one runtime artifact:
 
 - `gpu-services/` — Docker Compose stack for SenseVoice (STT) and CosyVoice
   (TTS) inference servers (`docker-compose.yml`, `healthcheck.sh`, model dirs).
-- `pi-orchestrator/` — Raspberry Pi deployment for the FastAPI orchestrator
-  (`vocalize.service`, `cloudflared-config.yml`, `deploy.sh`, `setup.sh`).
+- `orchestrator/` — Linux-host deployment for the FastAPI orchestrator,
+  with systemd + Cloudflare Tunnel wiring (`vocalize.service`,
+  `cloudflared-config.yml`, `deploy.sh`, `setup.sh`). Tested on Debian /
+  Ubuntu / Raspberry Pi OS; any modern Linux host with systemd works.
 
 Contrast with `scripts/` (maintainer-run utilities, not deployed) and `tools/`
 (release tooling, currently empty).
diff --git a/infra/pi-orchestrator/.env.template b/infra/orchestrator/.env.template
similarity index 100%
rename from infra/pi-orchestrator/.env.template
rename to infra/orchestrator/.env.template
diff --git a/infra/pi-orchestrator/.gitkeep b/infra/orchestrator/.gitkeep
similarity index 100%
rename from infra/pi-orchestrator/.gitkeep
rename to infra/orchestrator/.gitkeep
diff --git a/infra/pi-orchestrator/cloudflared-config.yml b/infra/orchestrator/cloudflared-config.yml
similarity index 93%
rename from infra/pi-orchestrator/cloudflared-config.yml
rename to infra/orchestrator/cloudflared-config.yml
index 92f4bbc..59c79c6 100644
--- a/infra/pi-orchestrator/cloudflared-config.yml
+++ b/infra/orchestrator/cloudflared-config.yml
@@ -1,6 +1,6 @@
-# Reference only — DO NOT deploy this file to /etc/cloudflared/ on Pi.
+# Reference only — DO NOT deploy this file to /etc/cloudflared/.
 #
-# Pi runs cloudflared via token-based service install:
+# Hosts run cloudflared via token-based service install:
 #   sudo cloudflared service install <TUNNEL_TOKEN>
 # That command embeds the tunnel id, credentials, and ingress config pulled from
 # the Cloudflare dashboard. No on-disk config.yml is required for runtime.
diff --git a/infra/orchestrator/deploy.sh b/infra/orchestrator/deploy.sh
new file mode 100755
index 0000000..ba84445
--- /dev/null
+++ b/infra/orchestrator/deploy.sh
@@ -0,0 +1,60 @@
+#!/usr/bin/env bash
+set -euo pipefail
+
+# deploy.sh — rsync VocalizeAI from your workstation to a remote Linux host
+# and restart the vocalize + cloudflared services.
+#
+# Usage:
+#   TARGET_HOST=<ip-or-hostname> ./deploy.sh
+#
+# Prerequisites:
+#   - SSH key auth to ${TARGET_HOST}
+#   - setup.sh has been run on the host at least once
+#   - cloudflared-config.yml has a real tunnel ID (not placeholder)
+#
+# Env vars (PI_HOST / PI_USER are accepted as legacy aliases of TARGET_HOST / TARGET_USER):
+#   TARGET_HOST    Remote host IP or hostname (required).
+#   TARGET_USER    SSH user on the remote host (default: current user).
+#   VOCALIZE_HOME  Remote project directory (default: /home/${TARGET_USER}/vocalize).
+
+TARGET_HOST="${TARGET_HOST:-${PI_HOST:-}}"
+if [ -z "${TARGET_HOST}" ]; then
+  echo "ERROR: set TARGET_HOST (or legacy PI_HOST) to your remote host IP or hostname." >&2
+  exit 1
+fi
+TARGET_USER="${TARGET_USER:-${PI_USER:-$(whoami)}}"
+TARGET_PROJECT_DIR="${VOCALIZE_HOME:-/home/${TARGET_USER}/vocalize}"
+LOCAL_PROJECT_DIR="$(cd "$(dirname "$0")/../.." && pwd)"
+
+echo "=== Deploying VocalizeAI to ${TARGET_USER}@${TARGET_HOST} ==="
+
+echo "[1/5] Rsync project to remote host..."
+rsync -avz \
+  --exclude '.venv' \
+  --exclude '__pycache__' \
+  --exclude '*.pyc' \
+  --exclude '.git' \
+  --exclude 'node_modules' \
+  --exclude '.omc' \
+  --exclude '.planning' \
+  --exclude '.env' \
+  "${LOCAL_PROJECT_DIR}/" \
+  "${TARGET_USER}@${TARGET_HOST}:${TARGET_PROJECT_DIR}/"
+
+echo "[2/5] Install/update Python deps on remote host..."
+ssh "${TARGET_USER}@${TARGET_HOST}" \
+  "cd ${TARGET_PROJECT_DIR} && .venv/bin/pip install -e ."
+
+echo "[3/5] Restart vocalize service..."
+ssh "${TARGET_USER}@${TARGET_HOST}" \
+  "sudo systemctl restart vocalize"
+
+echo "[4/5] Restart cloudflared service..."
+ssh "${TARGET_USER}@${TARGET_HOST}" \
+  "sudo systemctl restart cloudflared"
+
+echo "[5/5] Check service status..."
+ssh "${TARGET_USER}@${TARGET_HOST}" \
+  "sudo systemctl status vocalize cloudflared --no-pager -l"
+
+echo "=== Deploy complete ==="
diff --git a/infra/orchestrator/setup.sh b/infra/orchestrator/setup.sh
new file mode 100755
index 0000000..67888ed
--- /dev/null
+++ b/infra/orchestrator/setup.sh
@@ -0,0 +1,122 @@
+#!/usr/bin/env bash
+set -euo pipefail
+
+# setup.sh — one-time setup for running VocalizeAI on a remote Linux host.
+#
+# Tested on Debian 12, Ubuntu 22.04 / 24.04, and Raspberry Pi OS (Bookworm).
+# Any modern Linux distro with systemd works.
+#
+# Usage:
+#   1. Run this script from your workstation: TARGET_HOST=<ip-or-hostname> ./setup.sh
+#   2. Follow the manual cloudflared token-install + service-start steps printed at the end.
+#
+# Prerequisites:
+#   - Remote host running a recent Linux release with systemd
+#   - SSH key auth: ssh ${TARGET_USER}@${TARGET_HOST}
+#   - cloudflared installed on the host (sudo snap install cloudflared OR apt install cloudflared)
+#
+# Env vars (PI_HOST / PI_USER are accepted as legacy aliases of TARGET_HOST / TARGET_USER):
+#   TARGET_HOST    Remote host IP or hostname (required).
+#   TARGET_USER    SSH user on the remote host (default: current user).
+#   VOCALIZE_HOME  Remote project directory (default: /home/${TARGET_USER}/vocalize).
+
+TARGET_HOST="${TARGET_HOST:-${PI_HOST:-}}"
+if [ -z "${TARGET_HOST}" ]; then
+  echo "ERROR: set TARGET_HOST (or legacy PI_HOST) to your remote host IP or hostname." >&2
+  exit 1
+fi
+TARGET_USER="${TARGET_USER:-${PI_USER:-$(whoami)}}"
+TARGET_PROJECT_DIR="${VOCALIZE_HOME:-/home/${TARGET_USER}/vocalize}"
+SCRIPT_DIR="$(cd "$(dirname "$0")" && pwd)"
+
+echo "=== VocalizeAI Orchestrator One-Time Setup ==="
+echo "Target: ${TARGET_USER}@${TARGET_HOST}:${TARGET_PROJECT_DIR}"
+echo ""
+
+# ---- Step 1: Rsync repo to remote host ----
+echo "[1/5] Copying project to remote host..."
+rsync -avz \
+  --exclude '.venv' \
+  --exclude '__pycache__' \
+  --exclude '*.pyc' \
+  --exclude '.git' \
+  --exclude 'node_modules' \
+  --exclude '.omc' \
+  --exclude '.planning' \
+  --exclude '.env' \
+  "${SCRIPT_DIR}/../../" \
+  "${TARGET_USER}@${TARGET_HOST}:${TARGET_PROJECT_DIR}/"
+
+# ---- Step 2: Create venv and install deps ----
+echo "[2/5] Creating Python venv on remote host..."
+ssh "${TARGET_USER}@${TARGET_HOST}" "python3 -m venv ${TARGET_PROJECT_DIR}/.venv"
+
+echo "[3/5] Installing Python dependencies..."
+ssh "${TARGET_USER}@${TARGET_HOST}" "${TARGET_PROJECT_DIR}/.venv/bin/pip install --upgrade pip"
+ssh "${TARGET_USER}@${TARGET_HOST}" "${TARGET_PROJECT_DIR}/.venv/bin/pip install -e ${TARGET_PROJECT_DIR}"
+
+# ---- Step 3: Create .env if not exists ----
+echo "[4/5] Setting up .env..."
+ssh "${TARGET_USER}@${TARGET_HOST}" "test -f ${TARGET_PROJECT_DIR}/.env || cp ${TARGET_PROJECT_DIR}/infra/orchestrator/.env.template ${TARGET_PROJECT_DIR}/.env"
+echo "  -> .env created from template (edit it with real values before starting the service)"
+
+# ---- Step 4: Install systemd service ----
+echo "[5/5] Installing systemd vocalize service..."
+
+VOCALIZE_SERVICE=$(cat <<'SERVICEEOF'
+[Unit]
+Description=VocalizeAI orchestrator service
+After=network-online.target
+Wants=network-online.target
+
+[Service]
+# Install path is /opt/vocalize by convention; override by editing this unit
+# or by running setup.sh with VOCALIZE_HOME=/your/path set.
+Type=simple
+User=vocalize
+WorkingDirectory=/opt/vocalize
+EnvironmentFile=/opt/vocalize/.env
+ExecStart=/opt/vocalize/.venv/bin/python -m vocalize.main
+Restart=always
+RestartSec=5
+
+[Install]
+WantedBy=multi-user.target
+SERVICEEOF
+)
+
+ssh "${TARGET_USER}@${TARGET_HOST}" "echo '${VOCALIZE_SERVICE}' | sudo tee /etc/systemd/system/vocalize.service > /dev/null"
+ssh "${TARGET_USER}@${TARGET_HOST}" "sudo systemctl daemon-reload"
+ssh "${TARGET_USER}@${TARGET_HOST}" "sudo systemctl enable vocalize"
+
+echo "  -> vocalize.service installed and enabled"
+
+echo ""
+echo "=== Setup complete ==="
+echo ""
+echo "=== MANUAL STEPS REQUIRED ==="
+echo ""
+echo "1. Edit .env on the remote host with real values:"
+echo "     ssh ${TARGET_USER}@${TARGET_HOST} 'nano ${TARGET_PROJECT_DIR}/.env'"
+echo ""
+echo "2. Install cloudflared service on the host using a tunnel token (token-based auth):"
+echo "   a. Get the connector token from the Cloudflare dashboard:"
+echo "        Zero Trust → Networks → Tunnels → your-tunnel-name → Configure"
+echo "        → Install and run a connector → copy the long token string"
+echo "   b. SSH to the host and install the service with that token:"
+echo "        ssh ${TARGET_USER}@${TARGET_HOST} 'sudo cloudflared service install <TUNNEL_TOKEN>'"
+echo "      The token embeds tunnel id, credentials, and ingress config — no on-disk"
+echo "      config.yml is needed. (If you ever want to inspect intended ingress, see"
+echo "      infra/orchestrator/cloudflared-config.yml; that file is reference-only.)"
+echo ""
+echo "3. Make sure Public Hostname routing is configured in the dashboard:"
+echo "        api.example.com → http://localhost:8080"
+echo ""
+echo "4. Start services:"
+echo "     ssh ${TARGET_USER}@${TARGET_HOST} 'sudo systemctl start vocalize cloudflared'"
+echo ""
+echo "5. Verify both services are running and the public URL responds:"
+echo "     ssh ${TARGET_USER}@${TARGET_HOST} 'sudo systemctl status vocalize cloudflared --no-pager'"
+echo "     curl -fsS https://api.example.com/health"
+echo ""
+echo "After completing manual steps, use deploy.sh for subsequent deploys."
diff --git a/infra/pi-orchestrator/vocalize.service b/infra/orchestrator/vocalize.service
similarity index 100%
rename from infra/pi-orchestrator/vocalize.service
rename to infra/orchestrator/vocalize.service
diff --git a/infra/pi-orchestrator/deploy.sh b/infra/pi-orchestrator/deploy.sh
deleted file mode 100755
index 31333f1..0000000
--- a/infra/pi-orchestrator/deploy.sh
+++ /dev/null
@@ -1,50 +0,0 @@
-#!/usr/bin/env bash
-set -euo pipefail
-
-# deploy.sh — rsync VocalizeAI from Mac to Raspberry Pi and restart the vocalize service.
-#
-# Usage:
-#   ./deploy.sh
-#
-# Prerequisites:
-#   - SSH key auth to Pi at ${PI_HOST}
-#   - setup.sh has been run on the Pi at least once
-#   - cloudflared-config.yml has a real tunnel ID (not placeholder)
-
-PI_HOST="${PI_HOST:?set to your Pi IP or hostname}"
-PI_USER="${PI_USER:-$(whoami)}"
-PI_PROJECT_DIR="${VOCALIZE_HOME:-/home/${PI_USER}/vocalize}"
-MAC_PROJECT_DIR="$(cd "$(dirname "$0")/../.." && pwd)"
-
-echo "=== Deploying VocalizeAI to Pi ==="
-
-echo "[1/4] Rsync project to Pi..."
-rsync -avz \
-  --exclude '.venv' \
-  --exclude '__pycache__' \
-  --exclude '*.pyc' \
-  --exclude '.git' \
-  --exclude 'node_modules' \
-  --exclude '.omc' \
-  --exclude '.planning' \
-  --exclude '.env' \
-  "${MAC_PROJECT_DIR}/" \
-  "${PI_USER}@${PI_HOST}:${PI_PROJECT_DIR}/"
-
-echo "[2/4] Install/update Python deps on Pi..."
-ssh "${PI_USER}@${PI_HOST}" \
-  "cd ${PI_PROJECT_DIR} && .venv/bin/pip install -e ."
-
-echo "[3/5] Restart vocalize service..."
-ssh "${PI_USER}@${PI_HOST}" \
-  "sudo systemctl restart vocalize"
-
-echo "[4/5] Restart cloudflared service..."
-ssh "${PI_USER}@${PI_HOST}" \
-  "sudo systemctl restart cloudflared"
-
-echo "[5/5] Check service status..."
-ssh "${PI_USER}@${PI_HOST}" \
-  "sudo systemctl status vocalize cloudflared --no-pager -l"
-
-echo "=== Deploy complete ==="
diff --git a/infra/pi-orchestrator/setup.sh b/infra/pi-orchestrator/setup.sh
deleted file mode 100755
index 99146cd..0000000
--- a/infra/pi-orchestrator/setup.sh
+++ /dev/null
@@ -1,109 +0,0 @@
-#!/usr/bin/env bash
-set -euo pipefail
-
-# setup.sh — one-time setup for running VocalizeAI on a Raspberry Pi.
-#
-# Usage:
-#   1. Run this script from the Mac: ./setup.sh
-#   2. Follow the manual cloudflared token-install + service-start steps printed at the end.
-#
-# Prerequisites:
-#   - Raspberry Pi running Ubuntu 24.04
-#   - SSH key auth: ssh <pi-user>@<PI_HOST>
-#   - cloudflared installed on Pi (sudo snap install cloudflared  OR  apt install cloudflared)
-
-PI_HOST="${PI_HOST:?set to your Pi IP or hostname}"
-PI_USER="${PI_USER:-$(whoami)}"
-PI_PROJECT_DIR="${VOCALIZE_HOME:-/home/${PI_USER}/vocalize}"
-SCRIPT_DIR="$(cd "$(dirname "$0")" && pwd)"
-
-echo "=== VocalizeAI Pi One-Time Setup ==="
-echo ""
-
-# ---- Step 1: Rsync repo to Pi ----
-echo "[1/5] Copying project to Pi..."
-rsync -avz \
-  --exclude '.venv' \
-  --exclude '__pycache__' \
-  --exclude '*.pyc' \
-  --exclude '.git' \
-  --exclude 'node_modules' \
-  --exclude '.omc' \
-  --exclude '.planning' \
-  --exclude '.env' \
-  "${SCRIPT_DIR}/../../" \
-  "${PI_USER}@${PI_HOST}:${PI_PROJECT_DIR}/"
-
-# ---- Step 2: Create venv and install deps ----
-echo "[2/5] Creating Python venv on Pi..."
-ssh "${PI_USER}@${PI_HOST}" "python3 -m venv ${PI_PROJECT_DIR}/.venv"
-
-echo "[3/5] Installing Python dependencies..."
-ssh "${PI_USER}@${PI_HOST}" "${PI_PROJECT_DIR}/.venv/bin/pip install --upgrade pip"
-ssh "${PI_USER}@${PI_HOST}" "${PI_PROJECT_DIR}/.venv/bin/pip install -e ${PI_PROJECT_DIR}"
-
-# ---- Step 3: Create .env if not exists ----
-echo "[4/5] Setting up .env..."
-ssh "${PI_USER}@${PI_HOST}" "test -f ${PI_PROJECT_DIR}/.env || cp ${PI_PROJECT_DIR}/infra/pi-orchestrator/.env.template ${PI_PROJECT_DIR}/.env"
-echo "  -> .env created from template (edit it with real values before starting the service)"
-
-# ---- Step 4: Install systemd service ----
-echo "[5/5] Installing systemd vocalize service..."
-
-VOCALIZE_SERVICE=$(cat <<'SERVICEEOF'
-[Unit]
-Description=VocalizeAI orchestrator service
-After=network-online.target
-Wants=network-online.target
-
-[Service]
-# Install path is /opt/vocalize by convention; override by editing this unit
-# or by running setup.sh with VOCALIZE_HOME=/your/path set.
-Type=simple
-User=vocalize
-WorkingDirectory=/opt/vocalize
-EnvironmentFile=/opt/vocalize/.env
-ExecStart=/opt/vocalize/.venv/bin/python -m vocalize.main
-Restart=always
-RestartSec=5
-
-[Install]
-WantedBy=multi-user.target
-SERVICEEOF
-)
-
-ssh "${PI_USER}@${PI_HOST}" "echo '${VOCALIZE_SERVICE}' | sudo tee /etc/systemd/system/vocalize.service > /dev/null"
-ssh "${PI_USER}@${PI_HOST}" "sudo systemctl daemon-reload"
-ssh "${PI_USER}@${PI_HOST}" "sudo systemctl enable vocalize"
-
-echo "  -> vocalize.service installed and enabled"
-
-echo ""
-echo "=== Setup complete ==="
-echo ""
-echo "=== MANUAL STEPS REQUIRED ==="
-echo ""
-echo "1. Edit .env on the Pi with real values:"
-echo "     ssh ${PI_USER}@${PI_HOST} 'nano ${PI_PROJECT_DIR}/.env'"
-echo ""
-echo "2. Install cloudflared service on Pi using a tunnel token (token-based auth):"
-echo "   a. Get the connector token from the Cloudflare dashboard:"
-echo "        Zero Trust → Networks → Tunnels → your-tunnel-name → Configure"
-echo "        → Install and run a connector → copy the long token string"
-echo "   b. SSH to Pi and install the service with that token:"
-echo "        ssh ${PI_USER}@${PI_HOST} 'sudo cloudflared service install <TUNNEL_TOKEN>'"
-echo "      The token embeds tunnel id, credentials, and ingress config — no on-disk"
-echo "      config.yml is needed. (If you ever want to inspect intended ingress, see"
-echo "      infra/pi-orchestrator/cloudflared-config.yml; that file is reference-only.)"
-echo ""
-echo "3. Make sure Public Hostname routing is configured in the dashboard:"
-echo "        api.example.com → http://localhost:8080"
-echo ""
-echo "4. Start services:"
-echo "     ssh ${PI_USER}@${PI_HOST} 'sudo systemctl start vocalize cloudflared'"
-echo ""
-echo "5. Verify both services are running and the public URL responds:"
-echo "     ssh ${PI_USER}@${PI_HOST} 'sudo systemctl status vocalize cloudflared --no-pager'"
-echo "     curl -fsS https://api.example.com/health"
-echo ""
-echo "After completing manual steps, use deploy.sh for subsequent deploys."
diff --git a/install/pi-install.sh b/install/install.sh
similarity index 93%
rename from install/pi-install.sh
rename to install/install.sh
index 511c54b..f360b2d 100755
--- a/install/pi-install.sh
+++ b/install/install.sh
@@ -6,10 +6,10 @@ IFS=$'\n\t'
 # VocalizeAI Pi installer
 #
 # Deploys the VocalizeAI orchestrator on a Raspberry Pi.
-# Wraps the existing infra/pi-orchestrator/ assets.
+# Wraps the existing infra/orchestrator/ assets.
 #
 # Usage:
-#   bash install/pi-install.sh [--dry-run] [--steps "1,3,5"] [--skip-tunnel] [--skip-gpu]
+#   bash install/install.sh [--dry-run] [--steps "1,3,5"] [--skip-tunnel] [--skip-gpu]
 #
 # Flags:
 #   --dry-run       Print planned actions without performing any mutations.
@@ -53,7 +53,7 @@ while [[ $# -gt 0 ]]; do
             ;;
         *)
             echo "Unknown flag: $1"
-            echo "Usage: bash install/pi-install.sh [--dry-run] [--steps '1,3,5'] [--skip-tunnel] [--skip-gpu]"
+            echo "Usage: bash install/install.sh [--dry-run] [--steps '1,3,5'] [--skip-tunnel] [--skip-gpu]"
             exit 1
             ;;
     esac
@@ -128,7 +128,7 @@ step3_gpu_services() {
     echo "[3/7] GPU services setup..."
     echo "  GPU services (SenseVoice STT + CosyVoice TTS) run on a separate host."
     echo "  Ensure GPU_HOST in /opt/vocalize/.env points to that host's Tailscale IP."
-    echo "  This installer does not configure the GPU host — see docs/deploy/pi.md for details."
+    echo "  This installer does not configure the GPU host — see docs/deploy/linux.md for details."
     echo "[3/7] Done (note only — no mutation performed)."
 }
 
@@ -147,7 +147,7 @@ step4_tailscale_check() {
         else
             echo "  WARNING: Tailscale not found. Install it with:"
             echo "    curl -fsSL https://tailscale.com/install.sh | sh && sudo tailscale up"
-            echo "  Tailscale is required for the Pi to reach the GPU host."
+            echo "  Tailscale is required for this host to reach the GPU host."
         fi
     fi
     echo "[4/7] Done."
@@ -167,7 +167,7 @@ step5_cloudflared_tunnel() {
     echo "    Zero Trust -> Networks -> Tunnels -> [your tunnel] -> Configure"
     echo "    -> Install and run a connector -> Copy the token"
     echo ""
-    echo "  Reference ingress shape (docs only): infra/pi-orchestrator/cloudflared-config.yml"
+    echo "  Reference ingress shape (docs only): infra/orchestrator/cloudflared-config.yml"
     echo "  (Actual ingress routing is configured in the Cloudflare dashboard, not in a file.)"
     if [ "$DRY_RUN" = true ]; then
         echo "[DRY] would run: sudo cloudflared service install <TUNNEL_TOKEN>  (requires manual token)"
@@ -183,9 +183,9 @@ step6_systemd_unit() {
     echo ""
     echo "[6/7] Installing systemd unit and environment file..."
 
-    SERVICE_SRC="${REPO_ROOT}/infra/pi-orchestrator/vocalize.service"
+    SERVICE_SRC="${REPO_ROOT}/infra/orchestrator/vocalize.service"
     SERVICE_DST="/etc/systemd/system/vocalize.service"
-    ENV_SRC="${REPO_ROOT}/infra/pi-orchestrator/.env.template"
+    ENV_SRC="${REPO_ROOT}/infra/orchestrator/.env.template"
     ENV_DST="${INSTALL_DIR}/.env"
 
     # Install vocalize.service
@@ -237,7 +237,7 @@ step7_start_and_smoke() {
 # Main dispatcher
 # ---------------------------------------------------------------------------
 
-echo "=== VocalizeAI Pi Installer ==="
+echo "=== VocalizeAI Linux Installer ==="
 if [ "$DRY_RUN" = true ]; then
     echo "(DRY RUN — no system changes will be made)"
 fi
@@ -272,7 +272,7 @@ for i in 1 2 3 4 5 6 7; do
 done
 
 echo ""
-echo "=== Pi install complete ==="
+echo "=== Install complete ==="
 echo "Steps run: ${STEPS_RUN[*]:-none}"
 if [ "$DRY_RUN" = true ]; then
     echo "(DRY RUN — re-run without --dry-run to apply changes)"
diff --git a/scripts/README.md b/scripts/README.md
index 3ecb754..12f6f16 100644
--- a/scripts/README.md
+++ b/scripts/README.md
@@ -7,8 +7,9 @@ Each script is standalone and invoked from the repo root. Current scripts:
 - `build-public-filelist.py` — builds the filtered file list for the
   `sync-private-to-public` skill (consumes `install/public-allowlist.md`
   and `.public-sync-deny`).
-- `stability-24h-driver.py` — drives the 24-hour Pi stability rehearsal
-  (Phase 4 DEPLOY-02 evidence harness).
+- `stability-24h-driver.py` — drives the 24-hour orchestrator stability
+  rehearsal (Phase 4 DEPLOY-02 evidence harness). Hardware-agnostic; the
+  reference run was executed against a Raspberry Pi orchestrator.
 The dead-code-scanner whitelist (`vulture-whitelist.py`) was relocated to
 `.tooling/vulture-whitelist.py` in Phase 6 (maintainer-only scan tooling lives
 under `.tooling/` and is excluded from the public mirror).