Discord Translation for Large Servers
Translating a 50-person server is a solved problem. Most bots handle it, and so does copy-paste. What follows is about the three things that stop working somewhere between a few hundred and a few thousand members, why they stop, and what Pigona does about each one.
1. Translation loses the thread
Almost every translation bot works one message at a time. A message arrives, it goes to a translation API on its own, a translation comes back. In a quiet channel that is fine, because most messages are self-contained.
Busy channels are not made of self-contained messages. They are made of replies, pronouns, sentence fragments, and callbacks to something said two minutes ago. Translated in isolation, "yeah, that one" carries no information in any language. The faster the channel, the higher the share of messages that only mean something relative to what came before - so this problem grows with your server rather than staying flat.
The part that is easy to get wrong
The obvious fix is to send the last N messages along as context. In a busy channel that quietly fails, because the last N messages are not N messages of conversation. They are bot replies, command output, other translations, and one-word reactions. Take the last ten raw messages out of an active channel and you might get three of actual human conversation.
Pigona's Smart Context Engine counts only qualifying messages - written by a human, long enough to carry meaning - and reads further back to find them, so a depth of ten means ten real contributions to the conversation, not simply the last ten things posted. Selection is plain code, not a model call: it is deterministic, it adds no second model round-trip, and it stays inside a single Discord history request rather than fanning out into several. It will read at most fifty messages looking for them, so a channel that is mostly bot output cannot turn one translation into an unbounded scan.
You get two dials. Depth is how many qualifying messages to gather - 5, 10 or 20. Minimum length is how long a message has to be before it is worth gathering context for at all, which keeps "ok" and "lol" from costing you anything.
Context that trips an injection filter is discarded wholesale rather than sanitised and passed through, and the per-message context fee is waived when that happens - no charge for history that was thrown away. On a large server this is not hypothetical: the more members you have, the likelier one of them is curious about what happens if they type instructions at your bot.
Smart Context Engine is included on every server plan, and it is off until you set a depth for a channel - a new channel starts without it. It then adds a small per-analyzed-message fee on top of the translation itself, so depth is a cost dial you turn on deliberately rather than one that runs by default.
2. Every translation is another message in your channel
This is the most common complaint about translation bots generally, and it is purely a scale problem. A bot that posts a public translation under every foreign-language message doubles the traffic in that channel. At twenty messages an hour nobody minds. At two thousand, the channel becomes unreadable for everyone, including the people who did not need translating.
The useful question is not whether to translate, it is where the output should land. Pigona gives you four answers, and you can mix them per channel:
- Relay Channels send translations to a separate channel instead of the source one. They post through a webhook, so they carry the original author's name and avatar rather than reading as bot output. The source channel stays exactly as noisy as it was.
- Reactive Translation is on demand and public. Someone reacts with a flag and Pigona replies in the channel, so everyone can read it. For a private translation, use /translate or the right-click menu - those answer only the person who used them.
- Translate Buttons are a small panel under a foreign-language message with two buttons: one translates into the channel's language, the other into your own. Either way only the person who clicked sees the result. It is the on-demand middle ground - nothing is translated until somebody asks for it. The panel does not linger: it removes itself after a minute by default, and each new panel replaces the previous one, so a burst of messages leaves one panel rather than twenty.
- Autonomous Translation with a lifetime. If you do want public in-channel replies, they can be set to delete themselves after anywhere from a minute to a day. The translation is there when it is useful and gone before it becomes scrollback.
3. Moderation goes blind in languages your team does not read
A small server usually has one shared language and one culture of what is acceptable. A large one does not. Once a meaningful share of your traffic is in languages nobody on your moderation team reads, that traffic is effectively unmoderated - not because anyone decided to allow it, but because nobody can see it.
Discord's own internationalization guidance for this amounts to recruiting bilingual moderators, which works right up until you need a fourth language at 3am. Automated word filters do not fill the gap either, since every list has to be rebuilt by hand per language.
Being straight about what Pigona does here: it makes the content legible, and that is all. Autonomous Translation on a public channel means a moderator reading in their own language sees what is actually being said, in the same place, at the same time as everyone else. Pigona is not a moderation product, it does not score toxicity, and it does not integrate with AutoMod. What it removes is the excuse that nobody could read it.
4. Capacity, so you can stop wondering
Pigona bills in Translation Credits rather than advertising an unlimited plan, which invites the obvious question of whether the allowance survives a real server. Here is the credit-to-translation conversion for short messages, which is what most Discord traffic is:
| Plan | Price | Translations / month | Per day |
|---|---|---|---|
| Community | 15€ | up to 50,000 | ~1,700 |
| Pro | 29€ | up to 100,000 | ~3,300 |
| Business | 45€ | up to 200,000 | ~6,700 |
Two things make those numbers larger than the credit totals suggest. Charges are based on the translation produced, not the message read, so the original text and everything Pigona sends along with it are our cost rather than yours. And a message already in the target language never reaches the model, so it costs nothing at all - on our production traffic that is roughly a third of everything the bot reads. So in a mostly-English channel with a handful of German speakers, you pay for the German that gets translated. The English costs you nothing, however much of it there is.
Those are ceilings and worth reading as such. They price every message at the 0.2 Credit minimum, which is the whole charge for anything up to about forty characters, with Smart Context off - the configuration a channel starts in. A real channel is a mix of lengths, and measured across our own production traffic a Community server lands nearer 29,000 translations a month than 50,000. Smart Context adds a small fee per analysed message on top, and only once you switch it on for a channel.
Neither figure counts repeated-message caching, which in practice makes both better. Full pricing details.
What Pigona does not do
Worth knowing before you install rather than forty minutes after:
- A message already in the channel's target language is left alone by default. No translation, no Translate Buttons, no error, on every plan. This surprises people, because the obvious way to test a new channel is to post in your own language, which is the one case guaranteed to look like a failure. Post in a second language instead. Anyone can also react with a flag emoji to translate a message on demand, which works regardless. There is one server setting that changes this: you can have the Translate Buttons offered on those messages anyway, so a member can read one in their own language. It is off unless you turn it on, and it never makes Pigona translate them into the language they are already in.
- No video or other audio. Pigona reads images, PDFs, text files and voice messages, not video or other audio files.
- Attachments are read only when someone asks. A member right-clicks the message and chooses Translate Attachment. Nothing posted in a channel is translated automatically.
- Autonomous Translation posts publicly by default. If you want translations kept out of the source channel, that is a Relay Channel or a delete timer, and it is a setting you have to turn on.
- Smart Context Engine does not run until you switch it on. It is included on every plan, including the free one, but a new channel is created with a context depth of none - so a server nobody has configured translates message by message, and pays no context fee. Turning it on is a deliberate choice, per channel.
- Relay Channels need webhook permission in the destination channel, since that is what makes relayed messages look native instead of like bot output.
Try it on the busiest channel you have
The free tier is enough to see whether the context handling is real on your own traffic, which is the only test that settles it. No card, and you can point it at a single channel to start.