Engineered for serious cross-language conversations.
From understanding, to speaking, to professional correction, to team sharing — full pipeline covered.
Bidirectional simul-interpret
Two audio streams (your mic + system audio) drive ASR + translation in parallel. Left/right panels show your side and theirs separately. Turn-taking is instant.
- ✓ Source text in 100ms, translation streams in parallel
- ✓ Mic + system audio independent ASR pipelines
- ✓ Final segment = LLM-corrected source + translation pair
Local + cloud hybrid ASR
Whisper-faster (local GPU/CPU), Whisper API, Aliyun ASR, Qwen3-ASR — choose per scenario. Local first for privacy; cloud for max precision.
30+ languages, any pair
Chinese, English, Japanese, Korean, German, French, Spanish, Russian, Portuguese, Arabic, Hindi, Vietnamese, Thai, Turkish, Polish — and more. Auto-detect or manual locale lock.
Glossary management
Custom brand names, person names, technical terms — three categories. Group-based 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 translation prompt. LLM understands your business.
Virtual microphone output
TTS output routes to a virtual microphone. Zoom / Teams / Discord receive translated native voice. No copy-paste, no software switching.
- Your mic muted (no audio sent)
- You type or speak in native
- LLM translates & refines
- TTS speaks into the virtual mic
- Zoom / Teams / Discord hears native voice
Compose Bar (simul-interpret)
Type in your native language, see translation preview, manually correct, then send only when 100% correct. Zero translation accidents.
- 1 Type in native language
- 2 Live translation preview as you type
- 3 Manual edit if needed
- 4 Send: TTS + virtual mic
Quick phrases
Pre-set common phrases (greetings, small talk, business asks). Send any phrase with one click — text + TTS go out together via the virtual mic.
History & export
Every session is recorded automatically. Search, re-translate, summary generation, Markdown / PDF / DOCX export.
LAN API sharing
Expose local engines (ASR / TTS / LLM) as OpenAI-compatible APIs to other devices on your LAN. Team shares one GPU machine — saves cost.
curl http://192.168.1.42:18000/v1/chat/completions \
-H "Authorization: Bearer $LAN_TOKEN" \
-d '{"model":"local-llama","messages":[...]}'Cloud platform presets
OpenAI, Anthropic, Aliyun, DeepSeek, SiliconFlow — paste your API key, the engine list auto-populates. Switch providers in one click; no JSON wrangling.
Local-first by design
Whisper-faster + local TTS run entirely on your GPU/CPU. Sensitive meetings can be 100% offline. Cloud APIs are strictly opt-in — and even then, requests go from your machine directly to the provider, never proxied through CrossMeet servers.
- Audio never leaves your machine in local mode
- Cloud requests go direct to provider; we never proxy
- License + telemetry are minimal; logs auto-purge in 1h
- Hardware gating: no NVIDIA? local features auto-disable
CrossLearn — learn languages by watching
Embedded video player with smart subtitles. Pause on hover, double-click any word to save into your vocab. AI-generated dictionary cards explain context, etymology, and usage — built for active learning, not passive scrolling.
- ✓ YouTube + Netflix + Bilibili (via Subtitle Bridge browser extension)
- ✓ Hover-pause + word lookup with dictionary card
- ✓ Double-click to save into your personal vocab notebook
- 1 Open a video page in browser, extension pushes subtitles to CrossMeet
- 2 Subtitles render with vocab highlighting in real time
- 3 Hover a word to pause + see dictionary card
- 4 Double-click → save to vocab notebook for review
How CrossMeet compares
| Feature | CrossMeet | Otter.ai | Krisp | Google Translate |
|---|---|---|---|---|
| Real-time bilingual | ✓ | ✓ | ✓ | — |
| Local processing (privacy) | ✓ | — | partial | — |
| Virtual microphone output | ✓ | — | — | — |
| Knowledge base RAG | ✓ | — | — | — |
| Glossary management | ✓ | limited | — | — |
| 30+ languages | ✓ | EN only | EN only | ✓ |
| One-time pricing | ✓ | — | — | — |