Char

Overview

Char is a privacy-first, open-source AI notepad for meetings. It captures audio directly from your system instead of sending a bot into your calls. While you take notes, Char listens and transcribes everything in real time. After the meeting ends, it combines your notes with the transcript to produce a structured summary saved as plain markdown files on your device. You pick your AI stack: run fully offline with local models, bring your own API keys, or use Char's managed cloud. It works with Zoom, Teams, Google Meet, Slack, and in-person meetings.

Key Features

Local-First Transcription: Captures both microphone and system audio directly from your computer. No meeting bots, no cloud dependency, no permissions needed.
AI-Powered Summaries: Combines your handwritten notes with the full transcript to generate structured, accurate meeting summaries automatically.
Bring Your Own AI Stack: Choose between local models (via LM Studio or Ollama), your own API keys (OpenAI, Anthropic, Deepgram), or Char's managed cloud service.
Plain Markdown Storage: All notes, transcripts, and summaries are saved as markdown files on your device. Compatible with Obsidian, Notion, VS Code, and any text editor.
Fully Offline Capable: Run the entire pipeline locally, including speech-to-text and LLM summarization, with zero data leaving your machine.
Multi-Language Support: Transcribes meetings in 90+ languages including English, Spanish, French, German, Japanese, and more.
Advanced Templates (Pro): Override default AI prompts with custom template variables and instructions tailored to your workflow.
Enterprise-Ready Compliance: Designed for HIPAA, SOC2, ITAR, and FedRAMP-friendly deployments with on-premises hosting options.

Pros

True Data Privacy: Audio, transcripts, and notes never leave your device unless you explicitly choose cloud processing. No hidden data collection or AI training on your content.
No Meeting Bots: Captures audio at the system level, so there is no awkward bot joining your calls and no need to request recording permissions from participants.
Open Source and Transparent: Fully open-source codebase (7,900+ GitHub stars) you can inspect, modify, or self-host. No vendor lock-in or proprietary formats.
Flexible AI Options: Switch between fully local, hybrid, or cloud-based AI processing depending on your privacy and performance needs.
Works With Any Meeting Platform: Compatible with Zoom, Microsoft Teams, Google Meet, Slack, and even in-person meetings without any setup per platform.
Free Core Product: Unlimited local transcription and AI summaries at no cost. You only pay if you want managed cloud features.

Cons

macOS Only (Currently): Windows and iOS versions are not yet available, limiting access for non-Mac users.
No Speaker Identification Yet: Unlike some competitors, Char does not currently label who said what in the transcript.
Manual Recording Management: Users must remember to start and stop recordings manually. There is no calendar-based auto-start/stop feature yet.
Local AI Requires Hardware: Running models fully offline needs a reasonably powerful machine. Older or lower-spec devices may struggle with on-device transcription and summarization.

Use Cases

Confidential Client Calls: Lawyers, consultants, and financial advisors can record and summarize sensitive client conversations without any data leaving their device, meeting strict compliance requirements like HIPAA and SOC2.
Healthcare Documentation: Physicians and therapists can transcribe patient sessions locally, generating accurate clinical notes faster while keeping protected health information fully on-device.
Team Standups and Sprint Reviews: Engineering and product teams can capture daily standups or sprint retrospectives, getting automatic summaries with action items without adding a bot to every meeting.
Academic Lectures and Research: Students and researchers can record lectures or research interviews, then get AI-generated summaries and searchable transcripts stored as portable markdown files.
Sales Discovery Calls: Sales professionals can focus on the conversation while Char captures everything. Post-call summaries highlight key objections, next steps, and follow-up items.
Air-Gapped Enterprise Environments: Organizations in defense, government, or critical infrastructure can deploy Char on-premises with fully local AI models, ensuring zero external data transmission.

Frequently Asked Questions

What is Char?

Char is an open-source AI notepad for meetings. It transcribes your meetings in real time and generates structured summaries by combining your notes with the transcript. It runs locally on your device, keeping your data private.

How does Char capture meeting audio without a bot?

Char captures audio directly from your system's microphone input and speaker output. It listens at the operating system level rather than joining your meeting as a participant. This means no bot appears in your call, and no meeting permissions are needed.

What meeting platforms does Char work with?

Char works with any application that produces audio on your computer, including Zoom, Microsoft Teams, Google Meet, Slack, and phone calls. It also works for in-person meetings by capturing microphone audio.

Can I use Char completely offline?

Yes. You can run Char fully offline by using local AI models through LM Studio or Ollama. Both speech-to-text and summarization happen on your machine with zero internet connection required.

What languages does Char support?

Char supports 90+ languages for transcription, including English, Spanish, French, German, Japanese, Korean, Chinese, and many more.

Where are my notes stored?

All notes, transcripts, and summaries are saved as plain markdown files on your local filesystem. You can open them with any text editor, Obsidian, Notion, or VS Code. Char never stores your data on its servers unless you opt into cloud features.

What operating systems does Char support?

Char currently supports macOS (both Apple Silicon and Intel). Linux support is available via AppImage and .deb packages. Windows and iOS versions are in development.