Skip to main content
Version: v5.5.0

Limits

Settings → Admin → Limits controls how much each user may store and run, and how much work the compute worker and providers take on at once. It appears on server installs only. As with access and feature settings, each setting shows its source, can be reset, and can be seeded with RUNTIME_SEED_JSON. The maximum upload size is under Features, and generation tuning is under TTS Providers.

Limits​

SettingKeyWhat it does
Storage per accountlimitAccountStorage, accountStorageMbOptional total for documents and audio. Off by default; enabling reveals 10,240 MB.
Storage per guestlimitGuestStorage, guestStorageMbSeparate guest total. Off by default; enabling reveals 1,024 MB.
Total guest storagelimitGuestStoragePool, guestStoragePoolMbShared pool for all guests. Off by default; enabling reveals 10,240 MB. Accounts are unaffected.
Compute limitscomputeLimitPoliciesPer-user, site, worker and provider limits for every kind of compute work.

Every limit is off by default on a self-hosted install. Turn them on when you run a public or shared instance.

The compute limits cover PDF layout analysis, live playback, audiobook export, document previews, DOCX conversion, account export and TTS generation. The tab groups them as Daily allowances (new TTS generation and new documents), Reading and playback, Audiobooks and Background work, with worker and provider throughput under Capacity. Each kind of work has its own switch, and its own:

  • start windows: how often one user, or the whole site, can start it;
  • active work: how many can run at once, which recovers automatically after a crash;
  • queue and concurrency: queue length, priority and parallel jobs per worker, plus CPU, model, FFmpeg, LibreOffice and archive capacity.

Provider limits cap concurrent requests and the request and character rates sent to each provider, across all worker replicas. Add named overrides for providers on a different service plan. A provider's 429 Retry-After response pauses requests to that provider. Max concurrent is the number of segments generated in parallel for live playback. The default of three keeps playback ahead of the listener; lower it for a server that can only handle one request at a time.

Daily character limits​

TTS usage is counted in characters sent to a provider, and only for audio that has to be generated. Replaying cached audio is always free.

  • Before each missing segment is generated, OpenReader checks the user's usage. A segment that starts below the limit is finished and counted even if it crosses the limit; the next one stops.
  • Playback can continue through audio that is already generated, and seeking still works.
  • Retries never count twice.

Separate per-user limits apply to signed-in users and guests. They reset at midnight UTC. Users see their usage in Settings → Account, and admins see it per user in Users.

Creation caps and client addresses​

Optional creation caps under Access limit new accounts per IP over a rolling period and new guests per IP over the guest idle period. IPv6 addresses share a /64 bucket; stored IP values are keyed hashes. Existing accounts keep working when a creation cap is reached.

The creation caps read the client address from request headers, so they only hold when a reverse proxy sets those headers itself. On Vercel nothing is needed. Elsewhere, put OpenReader behind a proxy that overwrites X-Real-IP with the connecting address, and don't expose the app port directly: a client that reaches it without the proxy can send any address it likes. OpenReader reads X-Real-IP, then CF-Connecting-IP, then the last X-Forwarded-For hop.

ProxySetting
nginxproxy_set_header X-Real-IP $remote_addr;
Caddyheader_up X-Real-IP {remote_host} inside reverse_proxy

The complete policy, with every value, is in examples/openreader-seed.json. An incomplete or unknown policy in a seed stops startup before anything is written.

Storage and daily files​

New documents uses the document_intake policy. It is off by default; enabling it reveals 10 daily files per guest and 100 per account. Each upload preparation, text import or URL import counts once, including failed imports. URL imports check the allowance before fetching. The allowance resets at midnight UTC and has no worker execution block. Saved policies gain this action's defaults.

Settings → Storage, the admin user list and storage caps share one SQL total: source document sizes plus generated segment audio, timing sidecars and exports. Shared parsed layouts, previews and plans do not count. The worker measures the affected document during generation, when a run finishes and after audio deletion; opening settings never scans storage. The document settings panel still measures one document to separate audio by voice/settings and show shared derived data.

Upload preparation checks the projected total, and registration checks again under a transaction so concurrent uploads cannot bypass a cap. New audio checks the latest measured total before each uncached segment, even with character limits off. Measurements are periodic and in-flight segments can finish, so an audio run can cross the byte threshold before stopping. Cached audio remains playable. Blocked actions link to Storage to free space; guests can create an account when sign-ups are open.

Run Repair storage usage in Maintenance after upgrading to backfill existing audio before enabling caps. Its first run covers all users; later runs repair users active in the last seven days. Failed reports leave the previous total until a later run repairs it. Audio left by a removed document still counts until its queued cleanup or Reclaim orphaned audio removes it.