Resound

Resound turns voice conversations into portable local memory. It records or imports audio, transcribes with local Whisper or an OpenAI-compatible provider, and writes files you own: Markdown, JSONL, VTT, SRT, summaries, action items, and a consent-aware manifest.

No hosted service, dashboard, or database is required. Strata and TotalRecall are optional sinks.

Quick start

Requirements: Node.js 20, 22, or 24 and pnpm 9 through 11.

corepack enable
pnpm install
pnpm build
pnpm test
pnpm cli mock "Engineering Standup"

Inspect the result:

pnpm cli sessions list
pnpm cli validate <session>
pnpm cli export <session> --format all

Record or transcribe

On macOS, route call audio through a virtual input such as BlackHole, then run:

pnpm cli doctor --mode local-capture
pnpm cli audio devices
pnpm cli record --title "Client Call" --system <device-index> --mic <device-index>

Or transcribe an existing recording:

pnpm cli transcribe ./meeting.m4a --title "Q3 Planning" --provider local-whisper

See recording and providers for device routing and model configuration.

Commands

Command	Purpose
doctor	Check recorder and transcriber readiness
audio devices	List macOS inputs
record	Capture and transcribe audio
transcribe <file>	Import an existing recording
mock <title>	Run the complete offline workflow
sessions list/show	Inspect stored sessions
validate <session>	Validate consent, manifest, and outputs
export <session>	Generate Markdown, JSONL, VTT, and SRT
summarize / action-items	Regenerate derived artifacts
sink <target> <session>	Send to stdout, folder, webhook, Strata, or TotalRecall

Run pnpm cli --help for all flags.

Discord bot

The /resound bot is designed to run on your own Linux server, VPS, NAS, or local machine. It joins Discord voice through a DAVE-aware receiver, records separate speaker tracks, transcribes them, and keeps the artifacts on your host. No ReSound-hosted service is involved.

The safest first run uses mock audio. Copy the environment template, add your Discord token and application ID, then:

cp .env.example .env
pnpm bot:register
pnpm bot:start

For real server-side voice capture, install the pinned Python receiver dependencies and set RESOUND_BOT_MODE=discord. Docker Compose installs the complete runtime automatically. See self-hosting for the Discord application permissions, production configuration, upgrades, backups, and an end-to-end acceptance check.

Only members with Manage Server or an explicit configured operator user/role can start sessions. The session owner and operators control recording. Transcript downloads are ephemeral by default; public channel delivery is opt-in.

Architecture

• apps/cli: command-line interface
• apps/bot: Discord controller
• packages/audio: capture and receiver adapters
• packages/core: sessions, consent, validation, and storage
• packages/transcribers: local and remote providers
• packages/exporters: transcript and derived artifacts
• packages/sinks: optional destinations
• packages/kujo: executable checks

See usage, architecture, consent, and the changelog.