RCS — Reddit Comment Saver

Self-hostable Go backend + Chrome/Brave extension that auto-saves Reddit comments (and the thread’s top post) from www.reddit.com and old.reddit.com, with screenshots, GLOB search, and optional Imgur export.

Quick local dev

1. Backend

git checkout dev
make backend

Opens http://127.0.0.1:8080. SQLite + screenshots live under ./data/.

To drop local seed/test rows (keeps real Reddit captures):

make prune-dev-data

Optional env vars:

Variable	Default	Purpose
RCS_ADDR	127.0.0.1:8080	Listen address
RCS_DATA_DIR	data	DB + screenshots directory
RCS_DB_PATH	$RCS_DATA_DIR/rcs.db	SQLite path
RCS_API_KEY	(empty)	If set, require X-API-Key on API routes
RCS_PUBLIC_BASE_URL	(empty)	Canonical public origin (e.g. https://rcs.example.com)
RCS_TRUST_PROXY	(off)	Set 1/true to honor X-Forwarded-Proto / X-Forwarded-Host
IMGUR_CLIENT_ID	(empty)	Enables Imgur export in the Web UI

2. Extension (Chrome / Brave)

1. Open chrome://extensions (or brave://extensions)
2. Enable Developer mode
3. Load unpacked → select the extension/ folder
4. Click the RCS icon → enter backend URL http://127.0.0.1:8080 (and API key if you set one)
5. Browse a Reddit thread — visible comments are captured automatically (deduped by comment id)

3. Search Web UI

Open http://127.0.0.1:8080:

• Filter by username (prefix match)
• Free text uses SQLite GLOB (* and ?). Example: *bla*bla*
• Screenshot thumbnails appear when captured; click to enlarge (lightbox)
• Each result can Export to Imgur (needs IMGUR_CLIENT_ID). After export: Open on Imgur + Copy URL. The Imgur URL is stored so the same image is not re-uploaded.

Docker / VM / LXC

docker compose up -d --build

Or build a binary:

make build   # writes ./bin/rcs
RCS_ADDR=0.0.0.0:8080 RCS_DATA_DIR=/var/lib/rcs ./bin/rcs

Example systemd unit:

[Unit]
Description=RCS Reddit Comment Saver
After=network.target

[Service]
ExecStart=/usr/local/bin/rcs
Environment=RCS_ADDR=0.0.0.0:8080
Environment=RCS_DATA_DIR=/var/lib/rcs
Environment=RCS_API_KEY=
Environment=RCS_PUBLIC_BASE_URL=
Environment=RCS_TRUST_PROXY=
Environment=IMGUR_CLIENT_ID=
Restart=on-failure

[Install]
WantedBy=multi-user.target

User crontab (watchdog)

Keeps bin/rcs running without systemd: every minute the watchdog checks a pidfile and starts the process if it is down.

make build
cp -n data/rcs.env.example data/rcs.env   # if not created yet
# edit data/rcs.env (RCS_ADDR, Pangolin URL, API key, Imgur, …)
make install-cron

That installs:

# RCS watchdog
* * * * * /absolute/path/to/RCS/scripts/rcs-watchdog.sh

Logs append to data/rcs.log. Check status with tail -f data/rcs.log or make test-api. Remove with make uninstall-cron.

Cron does not rebuild Go — re-run make build after pulling updates. Override paths via RCS_BIN, RCS_ENV_FILE, etc. if needed (see scripts/rcs-watchdog.sh).

When the backend is not on localhost, grant the extension host permission when saving the URL in the popup (Brave/Chrome will prompt).

Reverse proxy (Pangolin)

RCS is meant to sit privately behind a reverse proxy. Auth stays at the proxy — leave RCS_API_KEY empty when Pangolin alone is enough.

Pangolin protected subdomain

1. Run RCS so only Pangolin/Newt can reach it (e.g. RCS_ADDR=0.0.0.0:8080 on a private site).
2. Create a public resource such as https://rcs.example.com pointing at upstream http://127.0.0.1:8080 (or the container IP).
3. Enable Pangolin authentication on that resource (protected / SSO).
4. Start RCS with:

RCS_ADDR=0.0.0.0:8080 \
RCS_PUBLIC_BASE_URL=https://rcs.example.com \
RCS_TRUST_PROXY=1 \
./bin/rcs

5. In the same browser profile, open https://rcs.example.com once and complete Pangolin login (sets the session cookie).
6. In the RCS extension popup, set Backend URL to https://rcs.example.com and grant host permission.

The extension sends cookies (credentials: "include") on API calls, so Pangolin allows the request the same way the Web UI tab does. Other browser profiles / Incognito need their own login. If health checks fail with 401/403 or an HTML login page, open the Web UI, log in, and retry.

Serve RCS at the subdomain root (/), not a subpath — static assets and API paths are absolute from /.

API (extension / integrations)

• GET /api/health
• GET /api/comments/exists?id=t1_... — { exists, has_screenshot }
• GET /api/posts/exists?id=t3_...
• POST /api/comments — JSON body with text fields + optional screenshot_base64 (data URL or raw base64)
• POST /api/comments/{id}/screenshot — attach screenshot to an existing comment (backfill)
• POST /api/posts
• GET /api/search?q=&user=
• POST /api/comments/{id}/imgur
• GET /api/blocklist — blocked usernames (synced to extensions)
• POST /api/blocklist — { "username": "AutoModerator" }
• DELETE /api/blocklist/{username}
• PUT /api/blocklist — replace full list { "usernames": [...] }

Layout

backend/     Go server (SQLite, Web UI, Imgur)
extension/   Manifest V3 Chrome/Brave addon
data/        Local SQLite + screenshots (gitignored)