For people who can read English but can't always hear it

See the source. Speak your native. Make English no barrier in meetings.

ASR puts source text on screen in 100ms — you understand instantly. TTS sends translated speech back into Zoom / Teams via a virtual microphone — they hear native English. Free 300 min/month, no card required.

Download free Win 10/11 · 300 min/mo free Watch 90-sec demo

7-DAY FREE TRIAL NO CREDIT CARD LOCAL-FIRST PRIVACY

CROSSMEET / Meeting Mode · Bidirectional

LIVE

EN · Source whisper-faster · gpu

00:02.1 Hello, this is John from Acme.

00:05.8 Thanks for joining the call today.

00:09.2 Let's start with the Q3 update.

ZH · Translation back-fill claude-sonnet · stream

+0.7s 你好，我是 Acme 的 John。

+0.9s 感谢你今天参加会议。

+0.8s 先从 Q3 进度更新开始吧。

ASR100ms LLM~0.9s vMicon RAGacme-2026.pdf

Trusted by

Engineers, sales teams,
and international consultants

~100ms

ASR-to-screen latency

30+

Languages, any pair

4×

ASR engines (local + cloud)

Three scenarios

Listen · Speak · Send. One tool, three modes.

Built around how a non-native speaker actually uses a meeting — not a generic translator wrapper.

Listening

Source text appears the moment ASR catches it — read 80% instantly. Translation streams in within ~1s and back-fills, never blocking the next sentence.

~100ms on-screen · 30+ languages · parallel LLM back-fill

Trade use-case →

Speak via virtual microphone

Speak natively. CrossMeet translates and pipes the TTS audio through a virtual microphone — Zoom, Teams or Discord hear it as native voice.

9 TTS engines · seamless vMic into Zoom / Teams · auto-narrate

Streamer use-case →

Business meeting · Compose Bar

Type in your language. Watch the live translation preview, edit it to 100%, then send — text + TTS together. Zero translation accidents.

compose-bar manual correction · bidirectional simul-interpret · send only when 100% correct

Lawyer use-case →

vs. the alternatives

One product covers what most need two or three tools to do.

Sources: vendor public pricing pages, May 2026. Facts only — no marketing adjectives.

Capability	CrossMeet	Otter.ai	Wordly	Teams Premium
Real-time two-way simul-interpret	✓	—	✓	Within Teams only
Virtual-mic loopback (they hear native voice)	✓	—	—	—
Local engines (offline, full privacy)	✓	—	—	—
Cross-platform meetings (Zoom / Teams / Meet / Discord)	✓	✓	✓	Teams only
BYO API Key (you pay your LLM/ASR provider directly)	✓	—	—	—
Starting price	$0 (free tier)	$8.33/mo	~$75/h	$10/seat/mo + M365

5 of 6 capabilities exclusive to CrossMeet · in this segment we are alone

Why CrossMeet feels different

Two iron rules that change the entire feeling.

Most "real-time" translators wait for the LLM, then show both source and translation together. CrossMeet doesn't.

RULE 01

Source text always arrives first.

100ms

The instant Whisper / Qwen3-ASR produces a token, it hits the screen — never waits for the translation pipeline. You read 80% before the LLM is even called.

seg_update event LLM bypass on src

RULE 02

LLM never blocks ASR.

∞parallel

Multiple LLM calls run in parallel and back-fill into the source slot. The ASR stream stays smooth and continuous — translation tucks itself into the right place.

Per-sentence concurrency Final = LLM-corrected

t = 0ms t = 100ms t = 480ms t = 900ms ← LLM back-fill

ASR · Source (EN)

···

"Hello,"

"Hello, this is John from Acme."

✓ "Hello, this is John from Acme."

LLM · Translation (ZH)

···

—

您好，我是 Acme 公司的 John。

Capabilities

Engineered for serious cross-language conversations.

Bidirectional simul-interpret

Two audio streams (mic + system) run independent ASR + LLM pipelines. Left/right panels stay separate. Turn-taking is instant.

Local + cloud hybrid ASR

Whisper-faster (local GPU/CPU), Whisper API, Aliyun ASR, Qwen3-ASR — choose per scenario. Sensitive sessions stay offline.

30+ languages, any pair

Chinese, English, Japanese, Korean, German, French, Spanish, Russian, Arabic, Hindi, Vietnamese, Thai, Turkish, Polish — auto-detect or manual lock.

Glossary management

Brand names, person names, technical terms — three categories with group activation per meeting. Injected into LLM prompt and validated post-hoc.

Knowledge base RAG

Upload company materials, product docs, customer briefs. Vector retrieval injects context into the prompt — the LLM speaks your business, not generic AI.

Virtual microphone (vMic)

TTS routes through a virtual mic device. Zoom / Teams / Discord pick it up as native voice — no copy-paste, no software switching.

Compose Bar (simul-interpret)

Type, watch live translation, manually correct, then send. Text + TTS together. 100% accurate when it matters.

History & smart export

Every session auto-recorded. Search, re-translate, AI summary, Markdown / PDF / DOCX export — recap-ready.

LAN API sharing

Expose local engines (ASR / TTS / LLM) as OpenAI-compatible endpoints. One GPU box serves the whole team.

CrossLearn — learn by watching

Watch videos with smart subtitles. Hover-pause, double-click words to save, AI-generated dictionary cards explain context — built for active learning.

Browse all features in detail →

Pricing

Free to start. Lifetime to stay.

BYO API Key — the software adds zero call cost. Free 300 min/month, paid plans unlock virtual mic + local engines + Interpret mode.

FREE

$0 / forever

300 min/mo · cloud engines

Real-time + text + learning
1 device

MONTHLY

$13 $9 / mo

Full Pro · flexible

Virtual mic + local engines
Interpret mode + RAG
2 devices