Cipher Assistent

Cipher Assistent is a minimal open-source bridge that lets a Telegram bot talk to GitHub Copilot Chat inside VS Code.

It is intentionally small:

• no OpenAI API layer
• no trading modules
• no project-specific logic
• no external model server
• just Telegram, a local Python bot, and a VS Code extension

The goal is simple: make it easy for other people to run a private Telegram-to-VS-Code assistant on their own machine.

What it does

1. A user sends a message to a Telegram bot.
2. The Python bot receives the message.
3. The bot forwards the request to a local VS Code extension over 127.0.0.1.
4. The extension opens Copilot Chat in VS Code and submits the prompt.
5. Copilot writes the final answer into a local response file.
6. The extension watches that file and sends the response back to the Python bot.
7. The Python bot sends the answer back to Telegram.

Everything stays local except the normal Telegram and Copilot services you already use.

Architecture

Telegram
   ↓
Python bot
   ↓ HTTP on localhost
VS Code extension
   ↓
GitHub Copilot Chat
   ↓ writes response file
VS Code file watcher
   ↓ HTTP on localhost
Python bot
   ↓
Telegram

Project structure

cipher-assistent/
├─ bot/
│  ├─ .env.example
│  ├─ bot.py
│  └─ requirements.txt
├─ extension/
│  ├─ .vscodeignore
│  ├─ package.json
│  ├─ tsconfig.json
│  └─ src/
│     ├─ botClient.ts
│     ├─ copilotBridge.ts
│     ├─ extension.ts
│     ├─ fileWatcher.ts
│     ├─ httpServer.ts
│     └─ logger.ts
├─ .gitignore
└─ LICENSE

Requirements

On your machine

• Windows, macOS, or Linux
• Python 3.11+
• Node.js 18+
• VS Code 1.85+
• A Telegram bot token from @BotFather and chatID from @userinfobot
• GitHub Copilot and GitHub Copilot Chat enabled in VS Code

VS Code extensions

Install these in VS Code:

• GitHub Copilot
• GitHub Copilot Chat

Setup

1. Clone the repository

git clone <your-repo-url>
cd cipher-assistent

2. Configure the Telegram bot

Create bot/.env from the example file.

cp bot/.env.example bot/.env

Then fill in:

• TELEGRAM_TOKEN
• ALLOWED_CHAT_IDS
• optional port overrides if needed

3. Install Python dependencies

cd bot
python -m pip install -r requirements.txt

4. Install extension dependencies

cd ../extension
npm install
npm run compile

5. Open the repository in VS Code

Open the repository root in VS Code.

Then open the extension folder and run the extension in Extension Development Host mode if you want to package or debug it.

For a local workflow, you can also compile the extension and install it as a VSIX.

Running the system

Step 1. Start the Python bot

From the bot folder:

python bot.py

The bot starts:

• the Telegram polling loop
• a local callback server on 127.0.0.1

Step 2. Start the VS Code extension by pressing F5 in your VS code where bot.py runs

When the extension activates, it starts a local HTTP server and begins watching the response file.

The extension exposes two commands:

• Cipher Assistent: Start Bridge
• Cipher Assistent: Stop Bridge

By default the bridge auto-starts when the extension activates.

Step 3. Talk to your bot on Telegram

Send a normal text message to the Telegram bot.

If your chat ID is allowed, the message is forwarded into Copilot Chat and the answer comes back to Telegram.

Configuration

Python bot environment variables

Variable	Required	Default	Description
TELEGRAM_TOKEN	yes	-	Telegram bot token
ALLOWED_CHAT_IDS	yes	-	Comma-separated list of allowed chat IDs
BOT_SERVER_PORT	no	4000	Local callback server port used by the Python bot
EXT_SERVER_PORT	no	3000	Local HTTP port used by the VS Code extension
REQUEST_TIMEOUT_SECONDS	no	600	Max wait time for a Copilot answer
TYPING_INTERVAL_SECONDS	no	4	Telegram typing indicator refresh interval
RESPONSE_FILE_PATH	no	workspace-root cipher_response.txt	Explicit response file path override

VS Code settings

Setting	Default	Description
cipherAssistent.extensionServerPort	3000	Port for the extension HTTP server
cipherAssistent.botServerPort	4000	Port for the Python bot callback server
cipherAssistent.responseFilePath	empty	Optional absolute override for the response file path

Security model

This project is intentionally local-first.

• The Python bot only binds to 127.0.0.1
• The VS Code extension only binds to 127.0.0.1
• Telegram access is restricted by ALLOWED_CHAT_IDS
• No OpenAI-compatible API server is included
• No project-specific secrets are part of the repository

Still, you should treat this as a personal local bridge and not as a hardened internet-facing service.

Why this repository exists

The original system lived inside a much larger private workspace with trading modules, experiments, and extra bridges.

This repository strips all of that away and keeps only the reusable core:

• Telegram bot
• local callback server
• VS Code extension
• Copilot response file bridge

That makes it easier for others to understand, run, and adapt.

Limitations

• It depends on GitHub Copilot Chat being available in VS Code
• It is designed for a single local machine workflow
• It does not include authentication beyond Telegram chat allow-listing
• It does not include conversation history storage, you can set this up in a memory.md file if needed
• It does not include model routing or provider abstraction

Customization ideas

You can extend this starter with:

• per-user sessions
• chat history persistence
• slash commands
• file attachments
• approvals before sending prompts
• multi-workspace routing
• alternative response-file formats
• module section

License

MIT