Most tools mishandle Persian for a reason hidden in plain sight: it's written in a script that looks like Arabic, so a recognizer treats "Arabic script" as "Arabic language" — and Persian is not Arabic. They share a right-to-left Perso-Arabic alphabet, but Persian (Farsi) is Indo-Iranian, an entirely separate language family with its own grammar and vocabulary. On top of that, Persian carries taarof, an elaborate register of ritual courtesy that a flat literal translation strips out — sometimes inverting what was meant. And spoken Tehrani isn't the Persian people write, so even a clean transcription can read like the wrong language. Add the Persian-English mixing of diaspora and tech teams, and "supports Persian" on a feature list tells you almost nothing. Here's what actually decides whether a Persian meeting comes back usable.
Same script as Arabic, a completely different language
Persian is written right-to-left in a Perso-Arabic alphabet, which is why tools that key off the script reach for Arabic — and get the language wrong. Persian is Indo-Iranian, related to languages across South and Central Asia, not to Semitic Arabic at all; the grammar, the verb system, the core vocabulary are unrelated. Persian even adds its own letters the Arabic alphabet doesn't have — پ, چ, ژ, گ — and a sentence routinely embeds left-to-right English words and Western numerals inside a right-to-left line, so the rendering has to handle both directions cleanly. A tool that lumps "Arabic script" into "Arabic" mishandles the words, the direction, and the mixed-script layout at once. For the genuinely Arabic case, which is a different problem, see Arabic meeting translation.
Taarof: the politeness that changes the meaning
Persian conversation runs on taarof — a register of ritual courtesy where offers are made to be politely declined, gratitude is layered, and intent is phrased indirectly. A host insisting "please, you first" or a guest's ceremonial refusal of an offer is a social form, not a literal position, and a flat word-for-word translation can render a ritual "no" as a real one — or strip a deferential, hedged request down to a blunt command it never was. This isn't mysterious or exotic; it's a formal politeness register, the way many languages have honorifics. But a translation that flattens it loses the indirection that carried the actual meaning, and the line comes back saying something the speaker didn't.
Spoken Tehrani isn't written Persian
Persian has a real gap between how it's written and how it's spoken — colloquial Tehrani contracts and reshapes words that the written form spells out in full. "I go" is mi-ravam on the page but miram in the room; vowels shift, endings collapse, and a recognizer trained on formal written Persian can miss the casual spoken form entirely. Layer on the Persian-English code-mixing common in diaspora and tech business — English nouns and verbs dropped into a Persian frame — and a tool that expects clean written Persian stumbles twice. (Worth noting too: Dari, the Persian variety spoken in Afghanistan, differs again in pronunciation and vocabulary, so "Persian" isn't a single uniform target.) A transcript has to follow the spoken room and still produce readable written Persian.
Why "supports Persian" isn't enough
A tool can list Persian, hear a sentence correctly, and still treat it as Arabic, render taarof as flat literal intent, miss the colloquial Tehrani forms, and leave Persian-English lines half-translated. The feature list won't tell you which. One real call will: does a native speaker read the captions and transcript and recognize how the room actually talked — right language, right direction, the politeness intact, the spoken forms caught? For why this pattern repeats across Asian and Middle Eastern languages, see real-time translation for remote teams.
How to do it with Sageio
- Add
bot@sageio.netto your Google Meet calendar invite. It joins on its own — no extension, nothing to install. - Each participant picks their caption language. A colleague in Tehran reads clean Persian in right-to-left script, a teammate abroad reads clean English — both from the same spoken Persian, at the same time. (Sageio translates into 20+ languages.)
- Everyone speaks naturally — Persian, the polite turns, the English mixing, all of it. Translated captions appear in about two seconds.
- Afterward, a searchable transcript and an AI summary arrive within about five minutes, shared at the host's discretion.
(Today this runs on Google Meet; Zoom and Microsoft Teams support is coming soon.)
How to test any tool in five minutes
Say a taarof-laden polite line first — a ceremonial offer or a deferential, hedged request — and check whether the translation keeps the indirection or flattens it into a blunt literal "no" or a flat command. Then say a normal Tehrani-colloquial sentence (miram rather than the written mi-ravam) and see whether the recognizer catches the spoken form instead of choking on it. Finally, say a mixed line with English words dropped in ("in feature-o next sprint deploy mikonim") and check two things at once: do the captions come back in right-to-left Persian script — not mistakenly Arabic — with the English and numbers sitting cleanly left-to-right inside the line? If the politeness flattens, the spoken forms get lost, or the direction and script break, the tool wasn't built for spoken Persian.
Is it private?
For anything that joins your meetings: Sageio doesn't use your meeting content to train AI models, and its AI vendors are contractually restricted from doing the same. Audio is processed in memory and discarded — only the text transcript and summary are kept, encrypted, in the region you choose (US, EU, or APAC). Enterprise customers can self-host the entire stack.
Frequently asked questions
Persian uses the same script as Arabic — aren't they basically the same? No. They share a right-to-left Perso-Arabic alphabet, but Persian (Farsi) is Indo-Iranian and Arabic is Semitic — separate language families with unrelated grammar and vocabulary. Persian even adds letters Arabic doesn't have (پ, چ, ژ, گ). A tool that treats "Arabic script" as "Arabic language" gets the words wrong even when the letters look familiar.
What is taarof, and why does it matter for translation? Taarof is Persian's ritual politeness register — offers made to be declined, indirect and deferential phrasing, layered courtesy. It's a formal register, not a quirk. A flat literal translation can render a ceremonial refusal as a real refusal, or strip a hedged request into a blunt command, so the line ends up saying something the speaker didn't mean.
Why is spoken Persian harder than written Persian? Colloquial Tehrani contracts and reshapes words the written form spells out — miram for written mi-ravam ("I go"). A recognizer tuned to formal written Persian can miss everyday spoken forms. Diaspora and tech speakers also mix English into a Persian frame, which adds a second thing for a one-language-per-sentence tool to trip over.
Does it handle Dari, the Afghan variety? Dari is the Persian spoken in Afghanistan and differs from Iranian Persian in pronunciation and some vocabulary, so "Persian" isn't a single uniform target. The honest check is the same: have a native speaker of the variety you actually use read the captions and transcript and confirm they recognize how the room talked.
What does it cost to try? Every plan starts with a free 60-minute trial, no credit card required. After that, Professional is $49/month and Teams is $99 per seat/month (annual billing includes 2 months free); Enterprise is custom-priced.
If your team works in Persian, the honest test is whether a native speaker reads the live captions and transcript and hears the actual meeting — right language, right direction, the politeness intact, the spoken forms caught, the English kept whole. Add the bot to your next call and let them judge.