## The Vision The idea is simple: you think of something worth sharing — a place, a person, a story, an idea, something you've learned — and you talk about it for two to three minutes.
That's it. No screens, no typing. Just a phone on a table with a good microphone, a conversation, and it all goes to the wiki.
Or when you are out for a walk in Nature.
Or when you are having a conversation with a friend or a colleague.
Anytime you use voice or even sound...why not record some birds singing or a train going by and see if the AI can translate that!
And Wiki equals proof of work. ---
## What You Need - **Telegram** (mobile and/or desktop — works on any platform) - **A transcription app** for your device (see the cross-platform section below) - **Python 3** installed on the computer where you run the bot script - **A shared folder** to keep everything synced (iCloud Drive on Mac, Google Drive or similar on other platforms) The bot script and the Telegram channel work on any device. The only thing that changes between platforms is which transcription app you use. --- ## Step 1: Create Your Telegram Bot You do this by talking to Telegram's BotFather. It's a conversation, not a form. 1. Open Telegram 2. Search for **@BotFather** 3. Start a chat 4. Send: `/newbot` 5. BotFather asks for a name — give it something meaningful, e.g. **Audio Stories** 6. BotFather asks for a username — it must end in `_bot`, e.g. **audiostories_bot** 7. BotFather gives you a **token** — a long string like `8556217280:AAGL5KnNVofXwKLVIbSr2XVJe3C94rxQYAE` **Keep this token safe.** Anyone with it can control your bot. --- ## Step 2: Create Your Telegram Channel 1. In Telegram, go to Menu → **New Channel** 2. Name it (e.g. **Audio Stories**) 3. Make it **Private** 4. Once created, tap the channel name at the top 5. Go to **Administrators** → **Add Administrator** 6. Search for your new bot (e.g. `@audiostories_bot`) 7. Add it as admin --- ## Step 3: Set Up MacWhisper Watch Folder MacWhisper can automatically transcribe any audio file that lands in a folder you choose. 1. Open **MacWhisper** on your Mac 2. Go to its settings 3. Set the **Watch Folder** to your Audio Stories folder, e.g.: `~/Library/Mobile Documents/com~apple~CloudDocs/HH/Audio_Stories/` Any `.ogg` file that appears there will be automatically transcribed. --- ## Step 4: Install the Python Package Open **Terminal** on your Mac and run: ```bash pip3 install python-telegram-bot ``` Wait for it to finish installing. --- ## Step 5: Save the Bot Script Download the `audio_stories_bot.py` script and save it to your HH folder in iCloud Drive: ```bash mv ~/Downloads/audio_stories_bot.py ~/Library/Mobile\ Documents/com~apple~CloudDocs/HH/ ``` This keeps it synced across all your devices. --- ## Step 6: Run the Bot ```bash cd ~/Library/Mobile\ Documents/com~apple~CloudDocs/HH/ python3 audio_stories_bot.py ``` You should see: ``` Audio Stories bot starting... Application started ``` Leave this running. The bot is now listening. --- ## How to Use It This is the daily workflow. It takes about 30 seconds to set up, then you just talk. ### 1. Set a Subject In your Audio Stories channel, type a subject followed by a colon: ``` Checkin: ``` or ``` Ideas_Wigtown: ``` or ``` DB_Checkin: ``` or ``` Theatre_of_the_Press: ``` Press send. The bot will reply: > ✅ Subject set: Ideas_Wigtown > Now send your voice message. ### 2. Send Your Voice Message Record and send a voice message in the same channel. The bot saves it automatically as: ``` Ideas_Wigtown_20260203_122526.ogg ``` ### 3. MacWhisper Does the Rest MacWhisper sees the new `.ogg` file in the watch folder and transcribes it automatically. You now have both the audio and the text. ### 4. Move to the Wiki When You're Ready When you're happy with the transcription, you choose what goes to the wiki. Nothing is pushed automatically — you stay in control. --- ## The Naming Convention The subject you type becomes the filename. Keep it clean and use underscores: | You type | File saves as | |---|---| | `Checkin:` | `Checkin_20260203_122526.ogg` | | `Ideas_Wigtown:` | `Ideas_Wigtown_20260203_122526.ogg` | | `DB_Checkin:` | `DB_Checkin_20260203_122526.ogg` | | `Theatre_of_the_Press:` | `Theatre_of_the_Press_20260203_122526.ogg` | | `Food_Sovereignty:` | `Food_Sovereignty_20260203_122526.ogg` | If you forget to set a subject, the file saves as `voice_20260203_122526.ogg` — no harm done, you can rename it later. --- ## Multiple Contributors Anyone can use the channel. David, Liam, others — they all follow the same process: 1. Type subject with a colon 2. Send voice message 3. Bot saves it with the subject name and a timestamp The `.txt` file created alongside each audio also records **who sent it** and **when**. --- ## Transcription Apps by Platform The Telegram bot and the subject naming system (`Checkin:`, `Ideas_Wigtown:`) works on **any platform**. It's only the transcription app that differs. ### macOS — MacWhisper (recommended) MacWhisper is the simplest option on Mac. Set a watch folder, and it transcribes everything automatically. See Step 3 above. ### Android - **FUTO Voice Input** — open source, privacy-focused, runs locally on your phone. Good if you don't want your audio going to the cloud. - **TalkNotes** — converts voice memos into organised text notes. Works in over 50 languages. Good for quick ideas on the go. ### Windows - **Vibe Transcribe** — free and open source. Runs locally, privacy-focused. Set up a watch folder the same way you would with MacWhisper. - **GoWhisper** — desktop app, runs locally. Affordable lifetime deal. Good for batch transcribing files. ### Web-based (any device with a browser) - **TurboScribe** — fast, supports over 98 languages, handles files up to 10 hours long. - **ListenMonster** — free, no sign-up required, processes multiple files at once. ### The Key Principle Whatever platform you're on, the workflow is the same: 1. Voice message goes into Telegram via the bot 2. Audio file lands in your shared folder 3. Your transcription app picks it up and converts it to text 4. You choose what goes to the wiki --- **Bot not responding?** Check that the script is still running in Terminal. If it stopped, restart it with `python3 audio_stories_bot.py`. **iCloud not syncing?** Open Terminal and run: `open ~/Library/Mobile\ Documents/com~apple~CloudDocs/` — this forces Finder to show the folder. **MacWhisper not transcribing?** Check that the Watch Folder in MacWhisper settings points to the correct `Audio_Stories` folder. --- ## The Prompt If you want to set this up yourself with Claude, here is the kind of prompt that gets you started: > "I want to create a Telegram bot that listens to a channel, receives voice messages, saves them with a subject-based filename, and drops them into a folder that MacWhisper can watch for automatic transcription. The subject is typed before the voice message, ending with a colon, like `Ideas_Wigtown:`. Can we build this step by step?" Then take it one step at a time. Don't rush ahead. Each step needs to be confirmed before moving to the next. --- *And if it all works first time? That earns a Binary Hug. 🤖🤗*