diff --git a/.env.example b/.env.example
index 1485412..36cc589 100644
--- a/.env.example
+++ b/.env.example
@@ -77,33 +77,23 @@ RUN_FS_MIGRATIONS=
 IMPORT_LIBRARY_DIR=
 IMPORT_LIBRARY_DIRS=
 
-# Heavy compute is always worker-backed.
-# App server calls compute-worker over HTTP for ONNX whisper alignment + PDF layout parsing.
-# Embedded/local (`pnpm dev` / `pnpm start` without COMPUTE_WORKER_URL) uses this same file.
-# `compute/worker/.env*` is only for standalone external worker deployments.
-# In single-container self-host setups, run compute-worker + NATS alongside the app and
-# point COMPUTE_WORKER_URL at that internal worker endpoint.
-# Required:
+# Compute
+# Embedded/local (default): leave COMPUTE_WORKER_URL empty.
+# External worker: set COMPUTE_WORKER_URL + COMPUTE_WORKER_TOKEN.
+# Details: docs/deploy/compute-worker
 # COMPUTE_WORKER_URL=http://localhost:8081
 # COMPUTE_WORKER_TOKEN=local-compute-token
-# Optional embedded worker stack controls (used by scripts/openreader-entrypoint.mjs):
-# If COMPUTE_WORKER_URL is unset, entrypoint defaults to starting embedded worker+NATS.
-# Set START_EMBEDDED_COMPUTE_WORKER=false to force external worker URL/token config.
-# START_EMBEDDED_COMPUTE_WORKER=
+# Optional embedded startup controls:
 # EMBEDDED_COMPUTE_WORKER_PORT=8081
 # EMBEDDED_NATS_PORT=4222
 # EMBEDDED_NATS_MONITOR_PORT=8222
 # EMBEDDED_NATS_STORE_DIR=docstore/nats/jetstream
 # NATS_URL=nats://127.0.0.1:4222
-# Optional shared compute timeouts used by:
-# - worker compute runtime
-# - worker client wait budgets
-# Worker-side concurrency is configured in compute-worker service via `COMPUTE_JOB_CONCURRENCY`.
+# Optional shared compute tuning:
 # COMPUTE_JOB_CONCURRENCY=1
 # COMPUTE_WHISPER_TIMEOUT_MS=30000
 # COMPUTE_PDF_TIMEOUT_MS=300000
 # COMPUTE_OP_STALE_MS=1800000
-# Worker mode requires worker-reachable shared object storage.
 
 # Optional Whisper ONNX base URL override (must contain all expected files)
 # WHISPER_MODEL_BASE_URL=https://huggingface.co/onnx-community/whisper-base_timestamped/resolve/main
diff --git a/compute/worker/.env.example b/compute/worker/.env.example
index f6188a6..3c38f5d 100644
--- a/compute/worker/.env.example
+++ b/compute/worker/.env.example
@@ -1,6 +1,6 @@
-# Standalone compute-worker env.
-# Used only when running compute-worker as a separate service/container.
-# Not used by app embedded worker startup via `pnpm dev` / `pnpm start`.
+# External compute-worker service env only.
+# Keep COMPUTE_WORKER_TOKEN and S3_* aligned with app env.
+# Details: docs/deploy/compute-worker
 
 # Compute worker bind
 # Platform note:
@@ -11,7 +11,7 @@
 # COMPUTE_LOG_FORMAT=pretty
 # COMPUTE_LOG_LEVEL=info
 
-# App <-> worker auth
+# Must match app env when app uses COMPUTE_WORKER_URL
 COMPUTE_WORKER_TOKEN=local-compute-token
 
 # NATS/JetStream
@@ -32,21 +32,17 @@ S3_SECRET_ACCESS_KEY=devsecret
 S3_ENDPOINT=http://host.docker.internal:8333
 S3_FORCE_PATH_STYLE=true
 
-# Queue + execution tuning
+# Optional tuning
 # COMPUTE_PREWARM_MODELS=true
 # COMPUTE_JOB_CONCURRENCY=1
 # COMPUTE_WHISPER_TIMEOUT_MS=30000
 # COMPUTE_PDF_TIMEOUT_MS=300000
-# Optional Whisper ONNX base URL override (must contain all expected files)
-# WHISPER_MODEL_BASE_URL=https://huggingface.co/onnx-community/whisper-base_timestamped/resolve/main
-# Optional PDF layout ONNX base URL override (must contain all expected files)
-# PDF_LAYOUT_MODEL_BASE_URL=https://huggingface.co/Bei0001/PP-DocLayoutV3-ONNX/resolve/main
 # COMPUTE_PDF_JOB_ATTEMPTS=1
 # COMPUTE_JOBS_STREAM_MAX_BYTES=268435456
-# JetStream stream size limit for op progress events (replay for SSE reconnect)
 # COMPUTE_EVENTS_STREAM_MAX_BYTES=134217728
 # COMPUTE_JOB_STATES_MAX_BYTES=67108864
 # COMPUTE_NATS_REPLICAS=1
-# Optional stale window for reusing in-flight opKey entries before forcing a new attempt
-# Default is max(30m, 4x max compute timeout); running jobs also refresh heartbeat every 5s
 # COMPUTE_OP_STALE_MS=1800000
+# Optional model mirrors
+# WHISPER_MODEL_BASE_URL=https://huggingface.co/onnx-community/whisper-base_timestamped/resolve/main
+# PDF_LAYOUT_MODEL_BASE_URL=https://huggingface.co/Bei0001/PP-DocLayoutV3-ONNX/resolve/main
diff --git a/docs-site/docs/deploy/local-development.md b/docs-site/docs/deploy/local-development.md
index 01dec45..14ca0a4 100644
--- a/docs-site/docs/deploy/local-development.md
+++ b/docs-site/docs/deploy/local-development.md
@@ -89,6 +89,41 @@ OpenReader currently pins `4.18` in CI and Docker builds while `4.19` compatibil
 
 </details>
 
+<details>
+<summary><strong>NATS Server <code>nats-server</code> (required for embedded compute mode)</strong></summary>
+
+If `COMPUTE_WORKER_URL` is unset, startup launches embedded compute worker + NATS, so `nats-server` must be available on host PATH.
+
+If you always use an external worker (`COMPUTE_WORKER_URL` set), this is not required.
+
+<Tabs groupId="local-dev-nats-os">
+<TabItem value="macos" label="macOS" default>
+
+```bash
+brew install nats-server
+nats-server -v
+```
+
+</TabItem>
+<TabItem value="linux" label="Linux">
+
+```bash
+# Linux amd64 example
+mkdir -p "$HOME/.local/bin"
+curl -fsSL -o /tmp/nats-server.zip \
+  https://github.com/nats-io/nats-server/releases/latest/download/nats-server-v2.12.1-linux-amd64.zip
+unzip -j /tmp/nats-server.zip '*/nats-server' -d /tmp
+install -m 0755 /tmp/nats-server "$HOME/.local/bin/nats-server"
+echo 'export PATH="$HOME/.local/bin:$PATH"' >> ~/.bashrc
+export PATH="$HOME/.local/bin:$PATH"
+nats-server -v
+```
+
+</TabItem>
+</Tabs>
+
+</details>
+
 <details>
 <summary><strong>LibreOffice (optional, for DOCX conversion)</strong></summary>
 
@@ -188,7 +223,6 @@ Default embedded worker flow (no external worker URL):
 ```env
 # Leave COMPUTE_WORKER_URL unset.
 # Entry point auto-starts embedded worker+NATS when available.
-START_EMBEDDED_COMPUTE_WORKER=
 ```
 
 External worker flow:
diff --git a/docs-site/docs/reference/environment-variables.md b/docs-site/docs/reference/environment-variables.md
index 190dc93..005ef32 100644
--- a/docs-site/docs/reference/environment-variables.md
+++ b/docs-site/docs/reference/environment-variables.md
@@ -55,7 +55,6 @@ For auth-enabled deployments, use **Settings → Admin** as the primary source o
 | `IMPORT_LIBRARY_DIRS` | Library import | unset | Set multiple roots (comma/colon/semicolon separated) |
 | `COMPUTE_WORKER_URL` | Heavy compute backend | unset | Set only for standalone external compute worker; leave unset for embedded worker startup |
 | `COMPUTE_WORKER_TOKEN` | Heavy compute backend | unset (auto-generated in embedded startup) | Required for standalone external compute worker auth; must match worker |
-| `START_EMBEDDED_COMPUTE_WORKER` | Heavy compute backend | auto (`true` when `COMPUTE_WORKER_URL` unset) | Set `false` to disable embedded worker startup and require external worker URL/token |
 | `EMBEDDED_COMPUTE_WORKER_PORT` | Heavy compute backend | `8081` | Override embedded worker bind port |
 | `EMBEDDED_NATS_PORT` | Heavy compute backend | `4222` | Override embedded NATS client port |
 | `EMBEDDED_NATS_MONITOR_PORT` | Heavy compute backend | `8222` | Override embedded NATS monitor port |
@@ -364,6 +363,7 @@ Multiple library roots for server library import.
 Base URL for standalone external compute worker mode.
 
 - Leave unset for embedded/local startup (`pnpm dev` / `pnpm start`) so entrypoint can start embedded worker+NATS.
+- Embedded startup requires `nats-server` available on host PATH.
 - Required only when using a standalone external worker service.
 - Example: `http://localhost:8081`
 
@@ -375,13 +375,6 @@ Bearer token for compute-worker auth.
 - Must match worker service `COMPUTE_WORKER_TOKEN`.
 - In embedded startup, entrypoint auto-generates one if unset.
 
-### START_EMBEDDED_COMPUTE_WORKER
-
-Controls whether entrypoint auto-starts embedded compute-worker + NATS.
-
-- Default behavior: enabled when `COMPUTE_WORKER_URL` is unset
-- Set `false` to disable embedded startup and require external worker URL/token
-
 ### EMBEDDED_COMPUTE_WORKER_PORT
 
 Embedded compute-worker HTTP port.
@@ -412,6 +405,7 @@ NATS connection URL used by compute worker runtime.
 
 - Embedded startup default: `nats://127.0.0.1:4222`
 - Standalone worker service: set in worker service env (`compute/worker/.env*` or platform env)
+- For embedded startup, this is optional; startup supplies the default value.
 
 ### COMPUTE_JOB_CONCURRENCY
 
diff --git a/scripts/openreader-entrypoint.mjs b/scripts/openreader-entrypoint.mjs
index 164cfd8..f21feb7 100644
--- a/scripts/openreader-entrypoint.mjs
+++ b/scripts/openreader-entrypoint.mjs
@@ -445,18 +445,13 @@ async function main() {
     const embeddedWorkerPort = Number.parseInt(withDefault(runtimeEnv.EMBEDDED_COMPUTE_WORKER_PORT, '8081'), 10);
     const embeddedNatsPort = Number.parseInt(withDefault(runtimeEnv.EMBEDDED_NATS_PORT, '4222'), 10);
     const embeddedNatsMonitorPort = Number.parseInt(withDefault(runtimeEnv.EMBEDDED_NATS_MONITOR_PORT, '8222'), 10);
-    const embeddedWorkerEnvRaw = runtimeEnv.START_EMBEDDED_COMPUTE_WORKER;
-    let shouldStartEmbeddedWorker = isTrue(
-      embeddedWorkerEnvRaw,
-      !Boolean(runtimeEnv.COMPUTE_WORKER_URL?.trim()),
-    );
+    const shouldStartEmbeddedWorker = !Boolean(runtimeEnv.COMPUTE_WORKER_URL?.trim());
 
     if (shouldStartEmbeddedWorker && !hasNatsBinary()) {
-      if (embeddedWorkerEnvRaw && isTrue(embeddedWorkerEnvRaw, true)) {
-        throw new Error('START_EMBEDDED_COMPUTE_WORKER=true but `nats-server` binary is not available in PATH.');
-      }
-      shouldStartEmbeddedWorker = false;
-      console.warn('`nats-server` binary not found; skipping embedded compute worker startup.');
+      throw new Error(
+        '`nats-server` binary is required when COMPUTE_WORKER_URL is unset. '
+        + 'Install nats-server or set COMPUTE_WORKER_URL and COMPUTE_WORKER_TOKEN for an external worker.',
+      );
     }
 
     if (shouldStartEmbeddedWorker) {