Why a Dev Built His Own Telegram Broadcast Tool Instead of Reusing Code

A developer who has shipped several dozen Telegram bots over the past year kept running into the same request: "Can you send a message to all our bot users?" After solving that problem five separate times with one-off scripts, small admin panels, or third-party SaaS tools he found too bloated, he built tgsend, a single dashboard for running Telegram broadcast campaigns across multiple bots. He shipped the first version in roughly one day and ran three live campaigns on launch day, including one to a list of about 3,000 chat IDs.

What happened

Detail	Fact
Tool name	tgsend
Build time	Roughly one day
Launch day campaigns	3 (2 own bots, 1 client)
Client list size	~3,000 chat IDs
Messages delivered	~2,000 (rest blocked or inactive)
Time to complete client send	~20 minutes
Send speed range	1 to 30 messages per second

The developer had built the same broadcast module repeatedly for different projects: as a throwaway script, as a small admin panel, and by handing off a bot token to a marketer. None of those options held up well over time. Scripts are one-and-done. Admin panels sprawl across projects. Sharing a bot token hands over more control than most clients should have. Full SaaS platforms come loaded with A/B testing, CRM logic, and auto-funnels that a coffee shop sending one newsletter per quarter simply does not need.

So he built tgsend instead: one workspace where you connect any number of bots, load recipient lists, compose messages, and fire off sends, without touching tokens, servers, or scripts directly.

How the tool actually works

The workflow has four steps:

1. Add a bot token. tgsend validates it against the official Telegram Bot API and then hides the token from the interface, so staff can run campaigns without having raw API access.
2. Load a recipient list as a file upload, a URL that auto-refreshes, or manual chat ID entry. Duplicates are removed automatically.
3. Compose the message. You can target the full list, specific users, or a random sample of N recipients. Attachments, buttons, photos, and multi-message sequences are all supported. A Rich Text mode uses Telegram’s newer Rich Message format, which allows images embedded in text, tables, and distinct content blocks inside one message.
4. Send or schedule. Before going live you can send a test to yourself. During the send, progress is visible in real time. You can pause and resume. When it finishes, you can export results to CSV or resend only to recipients who did not receive the message.

Why not just use an existing SaaS?

The developer is not dismissive of existing platforms. His argument is narrower: if your use case is daily funnels, segmentation, warmup sequences, and A/B tests, a full SaaS probably makes sense. But many small projects have a bot with somewhere between 200 and a few tens of thousands of users and need to send an announcement once a month or once a quarter. Paying a monthly SaaS fee for that volume is hard to justify.

His client base includes local businesses like cafes and retail shops. They do not care how the Telegram Bot API works. They need an employee to log in, write text, press “Test,” press “Send,” and leave. That is the gap tgsend is designed to fill. It is also the kind of workflow gap where a workflow automation layer can make a big operational difference for small teams.

How it shipped in one day

The developer credits two things. First, he used AI as a coding co-pilot, which he says genuinely accelerated the writing but did not replace his existing knowledge. Second, and more importantly, he already had the infrastructure in place: servers, a configured Telegram Bot API integration, message queues, and file storage. He did not give the AI access to his servers or VPS. The AI helped write code; the developer handled deployment and integration.

This mirrors what we see when helping clients with AI integration work: the speed gains are real, but they almost always come on top of a working technical foundation, not instead of one.

Our take

The core insight here is not the tool itself, it is the observation that drove building it. When you write the same module five times, the sixth time you should be shipping a service, not a script. That is sound engineering judgment, and the 20-minute, 2,000-message live send on day one is a better proof point than any staged demo.

What is worth noting for anyone running bots or automation for clients: keeping the bot token hidden from the end user while still letting them run campaigns is the right call from a security standpoint. Too many small bot setups hand over full API credentials to whoever happens to need a broadcast that week.

If you are managing Telegram bots across multiple clients or products, the pattern here, a single control plane that abstracts the token layer, is worth copying whether you build it yourself or find something off the shelf. For businesses thinking about how similar logic applies to their broader automation stack, you can see how we approach that kind of work in our client project write-ups.

The honest caveat: tgsend is described as “first for myself and clients, then we’ll see.” It is not a polished product yet. Evaluate it accordingly.

Building an automation like this? Most client workflows we ship run on Make (referral link, it supports our reporting). If you would rather have it built and monitored for you, that is our workflow automation service.

Source: Bing News · Telegram bots