Skip to content

telegram_bot_roadmap.md

Latest commit

 

History

History
65 lines (51 loc) · 3.99 KB

telegram_bot_roadmap.md

File metadata and controls

65 lines (51 loc) · 3.99 KB
Raw

Roadmap & Instructions: Automated Trading Chart Vision Bot

This document outlines the step-by-step roadmap for building an automated Telegram bot that receives a chart image, extracts trading parameters using Vision AI, calculates profit/loss percentages, and returns a strictly formatted message.

Target Output Format:

[ASSET] [ACTION] [ORDER_TYPE]
ENTRY: [VALUE]
SL: [VALUE]
TP: [VALUE]
Profit: +[X]% / Loss: -[Y]%

Phase 1: Telegram Bot Provisioning

Goal: Create the interface where users will send images and receive the signals.

  1. Create the Bot: Open Telegram and search for @BotFather. Use the /newbot command to create a new bot.
  2. Secure Credentials: Obtain the HTTP API Token provided by BotFather. This will be used by your backend server to communicate with Telegram.
  3. Configure Bot Settings: Set the bot's description and profile picture. Use BotFather to configure the commands menu (e.g., adding a /start command).

Phase 2: Vision AI Setup & Prompt Engineering

Goal: Configure a multimodal Large Language Model (like Google Gemini 1.5 Pro) to accurately "read" trading charts.

Reference Case Study: When the bot processes an image like BTCUSD_2026-07-09_10-34-20.png, it needs to identify the asset name in the top left ("Bitcoin / U.S. Dollar"), recognize the TradingView Short Position tool, and read the price labels on the right axis.

  • Grey Box (Middle): 63552.40 (Entry)
  • Red Box (Top): 63717.50 (Stop Loss)
  • Blue Box (Bottom): 60361.53 (Take Profit)

Instructions for AI Prompting:

  1. Enforce Structured Output: Do not let the AI write the final message directly. Instruct the AI to output only a strictly formatted JSON object containing the raw numbers and asset string.
  2. Define Visual Cues: Tell the AI in the prompt how to read the chart. For example: "The red shaded box represents the Stop Loss zone. The blue shaded box represents the Take Profit zone. The line separating them is the Entry price."
  3. Determine Order Type: Instruct the AI to deduce the order type. If the current price (e.g., 62913.22 in BTCUSD_2026-07-09_10-34-20.png) is below a short position entry, it is a "SELL LIMIT". If it's a long position above the current price, it's a "BUY LIMIT".

Phase 3: Backend Logic & Math (Avoiding AI Hallucinations)

Goal: Process the AI's data and accurately calculate the percentages. Never trust the AI to do the math.

  1. Parse the AI Response: Extract the JSON data containing Asset, Entry, SL, TP, and Action.
  2. Calculate Percentages (Instructions):
    • For a SELL signal:
      • Profit Percentage = ((Entry - TP) / Entry) * 100
      • Loss Percentage = ((SL - Entry) / Entry) * 100
    • For a BUY signal:
      • Profit Percentage = ((TP - Entry) / Entry) * 100
      • Loss Percentage = ((Entry - SL) / Entry) * 100
  3. Format the Data: Round the calculated percentages to two decimal places.

Phase 4: Message Formatting & Delivery

Goal: Assemble the final text and send it back to the user on Telegram.

  1. Construct the String: Map the parsed JSON values and calculated percentages to your exact requested template.
    • Example construction based on BTCUSD_2026-07-09_10-34-20.png: BTCUSD SELL LIMIT ENTRY: 63552.40 SL: 63717.50 TP: 60361.53
  2. API Delivery: Use your Telegram Bot Token to trigger the sendMessage endpoint, targeting the chat_id of the user who sent the picture.

Phase 5: Hosting and Deployment

Goal: Ensure the bot is responsive 24/7 without running on a personal computer.

  1. Choose a Cloud Provider: Select a backend hosting provider (e.g., AWS, Render, Heroku, or DigitalOcean).
  2. Set Environment Variables: Securely store your Telegram Token and Vision AI API keys in the cloud environment variables.
  3. Deploy as a Background Worker: Deploy your application so it constantly listens for incoming webhooks or polls the Telegram API for new messages.