Personal conversation copilot · macOS
A live record of both sides of the conversation.
dont-panic is a desktop app that captures your microphone and the other party’s audio as two labelled streams, transcribes both as you speak, and writes short first-person answers you can read in a glance. It runs on one Apple-Silicon Mac, for one person.
- 00:41ThemHow would you limit requests per user on a public API?
- 00:47YouGood question. Let me think about the storage first.
- 00:52ThemAnd what happens when the limit is (interim)
Answer · Suggest
I would start with a token bucket per user. It allows short bursts and is cheap to store.
What it does
- 01
Two sources, two labels
Your microphone is You. The other party’s audio (system audio) is Them. Each is captured and transcribed on its own, so the record always says who spoke.
- 02
Streaming transcript
Interim text appears while a person speaks and is replaced by final text when they finish. Each line carries a timestamp.
- 03
Short answers, on request
Ask writes a full answer; Suggest writes a short Teleprompter line; Screen AI reads a screenshot you trigger; Notes turns the conversation so far into a structured record.
- 04
Your knowledge base
Paste notes, a résumé or a job description, or take the text from a PDF. It grounds every Ask and Screen AI answer.
- 05
Swappable plugins
Audio capture, speech-to-text, turn detection and the LLM are each a plugin behind a stable Rust trait. Changing a provider is a setting, not a rewrite.
- 06
Notetaker mode
Record a meeting into a stored session instead, then summarise it with a template, search it, or chat over past sessions. Notetaker never answers live.
From audio to an answer
- Step 1
Capture two sources
Your microphone becomes You. System audio, from a virtual device such as BlackHole or from ScreenCaptureKit, becomes Them. A side with no audio is filled with silence so the timeline stays continuous.
- Step 2
Transcribe each side
Each side streams to the speech-to-text plugin you chose: Deepgram (a cloud service, needs your key) or Apple on-device speech recognition (no key, no server). Interim text arrives first; final text replaces it.
- Step 3
Detect the end of a turn
An energy-based turn detector watches Them and marks each End of utterance. Auto-fire uses it: Off, Shadow shows a ready cue, Live requests an answer.
- Step 4
Fire an Action
Ask, Suggest, Screen AI or Notes, from the interface or
Cmd+Shift+Enter. The request carries the recent transcript, the Rolling history and your Knowledge base as guidance notes. - Step 5
Stream the answer
The LLM you chose (Anthropic Claude, OpenAI or xAI Grok) streams the answer token by token into the Answer column, in short first-person lines.
- Step 6
Keep what you choose
In Notetaker mode the session is stored in a local SQLite database you can summarise, export, search and chat over. In Interview mode you save a transcript only when you choose to.
A personal tool, shared by request
dont-panic is built for one person on one Mac. There is no download page and no account. The source repository is private; ask for access.