Follow French meetings, lectures, and streams in English captions as they happen—no recording, no bot, just your Chrome tab.
French spoken at conversational pace—the standard estimate is around 350 syllables per minute in metropolitan France, faster in Montréal—packs grammatical gender, reflexive pronouns, and negation split across ne...pas into a tight stream. When you're sitting in on a product review with the Paris office, joining a webinar from a Québec software vendor, or watching a live panel from Lyon, you need captions that resolve agreement, reattach split verb phrases, and deliver meaning in English word order before the speaker finishes the next clause.
Translext captures the French audio playing in your browser tab, transcribes it, and generates English captions in real time. It runs entirely in the tab—no meeting invite for a bot, no file uploaded to a server, no transcript stored after you close the session. You read what's being said now, in English, while the meeting continues.
Every French noun carries masculine or feminine gender, and every adjective must match: nouveau projet, nouvelle stratégie. That agreement extends to past participles in compound tenses, pronouns, and determiners, often with no semantic weight—grammatical gender rarely maps to natural gender or meaning. When a speaker in Brussels describes une équipe technique or un système complet, the adjective ending signals which noun it modifies, a disambiguation English handles with word order instead. Translext parses gender and agreement in the French transcription, then produces English captions that drop the grammatical marking and rely on position and context. You see "technical team" and "complete system" without needing to track morphological concord across the sentence. Live translation will occasionally guess agreement wrong when audio quality drops or the speaker hesitates mid-phrase, but the English output remains intelligible because gender itself carries no lexical content to lose.
French pronunciation erases boundaries between words. Liaison adds consonants that don't appear in citation form: les enfants sounds like "lay-zahn-fahn," vous avez becomes "voo-zah-vay." Elision drops vowels before other vowels: je + ai contracts to j'ai, le + arbre to l'arbre. A listening tool must decide whether a spoken phrase is les amis or l'ami, c'est une or ses unes, without the whitespace that written French preserves. These phonetic merges happen constantly in metropolitan French and carry through Francophone Africa, though Québec French shifts some liaison patterns and inserts affricate consonants—tu becomes "tsu"—that further blur word edges. Translext transcribes the phonetic stream, segments it back into lexical units, and renders the phrase in English where spaces between words are stable. When you're following a supplier call from Dakar or a training video recorded in Toulouse, you see "the friends" or "his friends" in the caption without manually parsing sandhi.
French maintains a second-person singular informal tu and a formal vous that English collapsed centuries ago into universal "you." In French business contexts—especially finance, law, and corporate settings in France, Belgium, Switzerland, and Luxembourg—colleagues may vouvoie each other for years, even when working on the same team daily. The choice signals hierarchy, distance, or deliberate respect, and switching from vous to tu mid-meeting is a social event. Translext transcribes both forms but translates them identically to "you" because English has no morphological equivalent. The captions won't tell you whether the Paris compliance officer just used tu or vous with the Geneva auditor, but you'll understand the substance of the question being asked. If register matters to your context—if you're following HR mediation or executive negotiation where formality shifts carry weight—watch for softening phrases like "si vous permettez" or "tu vois" that mark tone even after the pronoun itself becomes invisible in English.
Translext transcribes metropolitan French, Québec French, and Francophone African varieties, but accuracy varies with accent density and recording quality. Québec French shortens high vowels, affricates /t/ and /d/ before /i/ and /y/, and uses vocabulary like "courriel" where France says "e-mail." Ivorian and Senegalese French often layer tone patterns from Baoulé or Wolof substrate languages. The transcription engine recognizes these features in clear audio; heavy regional accent or low bitrate will produce more errors. You'll see mistakes in the captions—wrong verb forms, dropped negations—but the general meaning usually comes through.
Yes. Translext captures audio from any tab in Chrome—a YouTube live stream, Google Meet, Zoom running in browser, Microsoft Teams web app, or a video playing on a conference platform. Click the extension icon, select the tab, pick French as the source language and English as the target, and captions appear in the extension panel. The tab keeps playing; the extension listens and transcribes in real time. It does not join the meeting as a participant, does not record the session, and does not save a transcript after you stop. No other browsers—Firefox, Safari, Edge—are supported today.
French speakers borrow English terms constantly, especially in technology, finance, and management: "le cloud," "IA" for intelligence artificielle, "faire un brief," "pitcher." Translext transcribes these as they're spoken, often keeping the English term in the French transcription, then passes it into the English caption unchanged or lightly adapted. Occasionally the transcription will mishear an English loanword as a French phrase—"deadline" heard as "dead ligne"—and the English caption will look strange until context resolves it. The system does not flag code-switching explicitly; you just see the English word appear in both the transcription and the translation.
Captions lag behind the live audio by a few seconds because the transcription engine waits for enough context to close a phrase boundary, and the translation step adds another brief delay. French's verb-final negation—ne...pas wrapping the verb—and noun phrases with trailing prepositional qualifiers mean the system often needs the end of a clause before it can render accurate English word order. If two people talk over each other, the transcription will capture fragments from both voices mixed together, and the English caption will be garbled. Translext does not separate speakers or label who said what; it treats all audio in the tab as a single stream.
Spoken French routinely drops ne and signals negation with pas, jamais, or rien alone: "je sais pas" instead of "je ne sais pas." Translext transcribes what it hears, so the French transcription will often show the colloquial form without ne. The translation engine infers negation from pas or jamais and produces "I don't know" or "never" in the English captions correctly most of the time. Occasionally, in fast or unclear speech, a standalone pas is missed or mistranscribed as a different word, and the English caption will lose the negation entirely. You'll see an affirmative sentence where the speaker meant negative. Check context if a caption seems to contradict what came before.
Live captions and translated audio for anything playing in your browser tab. No bots to invite, nothing to upload.