ChatGPT-Telegram-Workers is a project designed to deploy your own ChatGPT Telegram bot easily. It is optimized for serverless deployment, specifically targeting Cloudflare Workers, allowing for a single-file 'copy-paste' setup without needing a local development environment, a domain, or a dedicated server.

Key features include:

• Serverless Deployment: Supports Cloudflare Workers, Vercel, and Docker.
• Multi-AI Provider Support: Compatible with OpenAI, Azure OpenAI, Cloudflare AI, Cohere, Anthropic, Mistral, and more.
• Bot Management: Supports multiple Telegram bots, streaming output, and multi-language support.
• Interactive UI: Uses InlineKeyboards for model switching and supports custom commands for switching models or bot presets.
• Extensibility: Includes a plugin system for custom functionality and supports text-to-image generation.

If you are reusing a wrangler.toml configuration file from Cloudflare Workers, you can synchronize those environment variables to Vercel by running pnpm run vercel:syncenv.

Important: After modifying environment variables in the Vercel console, you must perform a redeployment for the changes to take effect.

Because forked repositories do not automatically sync with the upstream repository, you must manually sync them to trigger deployments. To automate this synchronization, create a file at .github/workflows/sync.yml in your repository and use the following configuration.

Note: This implementation is provided for reference only and has not been tested.

name: Sync
on:
  schedule:
    - cron: '0 0 * * *'
jobs:
    sync:
        runs-on: ubuntu-latest
        steps:
        - name: Sync
            uses: repo-sync/github-sync@v2
            with:
            source_repo: 'https://github.com/TBXark/ChatGPT-Telegram-Workers'
            target_repo: 'Fill in your repository address'
            github_token: ${{ secrets.GITHUB_TOKEN }} 
            source_branch: 'master'

You can use the Vercel deployment button to automatically clone the repository and deploy it to your Vercel account. This process will prompt you to configure necessary environment variables such as UPSTASH_REDIS_REST_URL, UPSTASH_REDIS_REST_TOKEN, and TELEGRAM_AVAILABLE_TOKENS.

[![Deploy with Vercel](https://vercel.com/button)](https://vercel.com/new/clone?repository-url=https%3A%2F%2Fgithub.com%2FTBXark%2FChatGPT-Telegram-Workers&env=UPSTASH_REDIS_REST_URL,UPSTASH_REDIS_REST_TOKEN,TELEGRAM_AVAILABLE_TOKENS&project-name=chatgpt-telegram-workers&repository-name=ChatGPT-Telegram-Workers&demo-title=ChatGPT-Telegram-Workers&demo-description=Deploy%20your%20own%20Telegram%20ChatGPT%20bot%20on%20Cloudflare%20Workers%20with%20ease.&demo-url=https%3A%2F%2Fchatgpt-telegram-workers.vercel.app)

To enable the automated deployment workflow, you must add specific secrets to your forked repository's GitHub settings and enable Actions.

1. Add Secrets

Navigate to your repository's Settings -> Secrets and add the following:

2. Enable Actions

Navigate to Settings -> Actions in your GitHub repository and ensure Actions are enabled.

To enable automated deployment via GitHub Actions, you must create a Cloudflare API Token with specific permissions:

1. Log in to your Cloudflare account and navigate to My Profile.
2. Select API Tokens from the left menu.
3. Click Create Token.
4. Select the Edit Cloudflare Workers template.
5. Under Zone Resources, select the appropriate zone.
6. Under Account Resources, select the appropriate account.
7. Click Create Token.

Note: Keep this token secure and do not share it. It is required for the GitHub Action to deploy your workers.

After deploying your Worker to Cloudflare, you must trigger the initialization endpoint to bind the Telegram webhook and configure all necessary commands.

Access the following URL in your browser: https://<your-worker-name>.<your-username>.workers.dev/init

You can run the project locally using pnpm via two methods:

Method 1: Direct start

pnpm install
pnpm run start:local

Method 2: Build and start with custom config paths Use this method to explicitly define the CONFIG_PATH and TOML_PATH environment variables.

pnpm install
CONFIG_PATH=./config.json TOML_PATH=./wrangler.toml pnpm run start:dist

You can deploy the project to Vercel using the provided deployment button. Note that the Vercel deployment implementation is currently experimental and located in /packages/app/vercel. While it allows for basic functional testing, it does not guarantee that all features will work correctly.

[![Deploy with Vercel](https://vercel.com/button)](https://vercel.com/new/clone?repository-url=https%3A%2F%2Fgithub.com%2FTBXark%2FChatGPT-Telegram-Workers&env=UPSTASH_REDIS_REST_URL,UPSTASH_REDIS_REST_TOKEN,TELEGRAM_AVAILABLE_TOKENS&project-name=chatgpt-telegram-workers&repository-name=ChatGPT-Telegram-Workers&demo-title=ChatGPT-Telegram-Workers&demo-description=Deploy%20your%20own%20Telegram%20ChatGPT%20bot%20on%20Cloudflare%20Workers%20with%20ease.&demo-url=https%3A%2F%2Fchatgpt-telegram-workers.vercel.app)