API is free, real costs are hosting and tooling. Stars-to-USD math, four use case scenarios, and an honest no-code builder comparison.
By Veracity Media Network · Published 2026-04-19
This page breaks down each of those numbers, explains where the confusion comes from, and compares the five no-code builders that actually support Telegram today.
Almost every page ranking for "telegram bot api free" says the API is free and leaves you to figure out the rest. That framing is technically correct but misleading, because it leaves four costs silent: hosting, AI APIs if your bot uses them, Paid Broadcasts if you need them, and the builder or developer you pay to actually assemble the bot.
Here is what the Telegram Bot API genuinely includes at zero cost:
The Telegram Bot API has been free since launch in 2015 with no per-message fees and no cap on the number of bots per account. [1]
That is an unusually generous free tier for an API with 950 million reachable users. The catch is that the API only handles Telegram's side of the conversation. Everything your bot actually does, from parsing messages to calling an LLM to storing state to responding with context, runs on infrastructure you pay for.
Three categories of real cost show up for almost every bot:
Hosting. Your bot runs somewhere. A lightweight notification bot handles a few hundred users on a $5/month VPS at Hetzner or DigitalOcean. A moderately active bot serving 1,000-10,000 users typically needs $15-40/month. High-traffic bots with thousands of concurrent users may run $50-200/month on larger cloud instances.
AI APIs. If your bot uses an LLM to understand or respond to messages, you pay per token to OpenAI, Anthropic, or another model provider. For customer support bots handling moderate volume, this usually runs $10-50/month on cost-efficient models (GPT-4o mini, Claude Haiku). Bots using frontier models for every interaction can run hundreds per month.
Database. If your bot remembers anything about users between sessions, you need storage. Free tiers from Supabase, Neon, or PlanetScale cover most bot workloads. SQLite on your VPS is free. You rarely pay anything here until you scale past tens of thousands of active users.
None of this is Telegram's cost. It is what a bot needs to exist at all. The "free API" marketing hides the fact that you are comparing Telegram to other platforms where messaging is the thing you pay for (WhatsApp at $0.005-$0.08 per conversation), and where hosting and compute are shared concerns you pay for in both cases.
The Paid Broadcasts feature costs "0.1 Telegram Stars per message above 30 per second." Every competitor page we checked quotes this number and stops there, leaving readers to guess what 0.1 Stars actually costs in dollars. That is an easy gap to fill, and it is where this page delivers something the others do not.
Telegram Stars are not a standalone cryptocurrency. They are a balance inside Telegram's system, convertible to Toncoin (TON) at a fixed rate of 200 Stars = 1 TON. TON trades on public crypto markets, and its USD price determines what Stars are worth.
0.010-0.025 USD per Telegram Star
At April 2026 TON market prices ($2.00-$5.00 USD), one Telegram Star is worth approximately $0.010-$0.025 USD. The 0.1 Stars charged per Paid Broadcast message above the 30 msg/sec free tier costs roughly $0.001-$0.0025 per message. The 10,000 Stars activation threshold for Paid Broadcasts costs $100-$250 USD upfront depending on TON's spot price.
Verified as of April 19, 2026:
| TON/USD price | 1 Star (USD) | 0.1 Star / Paid Broadcast message (USD) | 10,000 Stars activation threshold (USD) |
|---|---|---|---|
| $2.00 | $0.010 | $0.0010 | $100 |
| $2.50 | $0.0125 | $0.00125 | $125 |
| $3.00 | $0.015 | $0.0015 | $150 |
| $4.00 | $0.020 | $0.0020 | $200 |
| $5.00 | $0.025 | $0.0025 | $250 |
TON has ranged between approximately $1.80 and $5.00 since Stars launched in mid-2024, so a realistic budget range for a bot using Paid Broadcasts is:
This matters for two use cases specifically:
High-volume broadcast channels. A news bot pushing to 50,000 subscribers hits the 30 msg/sec cap almost instantly during a broadcast. Sending the 50,000 messages at the paid rate costs roughly 5,000 Stars above the free tier, or $50-125 per broadcast at typical TON prices. For a bot sending a few broadcasts per day to a large list, this adds up to $500-1,500/month.
Bots that need burst capacity. A crypto trading bot or a time-sensitive alerting bot may rarely hit the limit but needs headroom when it does. Paying for the 10,000 Stars activation is essentially an insurance policy for events that cannot wait.
For the vast majority of bots (customer support, notifications to small-to-medium user bases, personal assistants, group moderation, e-commerce order updates), the 30 msg/sec free tier is more than enough, and the Stars question is irrelevant. The feature exists for high-volume broadcast use cases that are a narrow slice of the total bot population.
Generic price ranges are less useful than concrete scenarios. Here is what four common bot archetypes actually cost per month at typical scale.
Example use case: Internal team alerts when a deployment completes, when a monitoring threshold trips, or when a customer support ticket comes in.
| Component | Typical cost |
|---|---|
| Telegram API | $0 |
| Hosting (minimal VPS or serverless) | $0-5/month |
| AI APIs | $0 (no LLM needed) |
| Database | $0 (free tier or none) |
| Total | $0-5/month |
These bots rarely need more than a few dozen messages per day. Free-tier serverless (Vercel, Cloudflare Workers, Railway) handles them entirely. Our research database shows 46 of 51 conversational AI tools we have audited (90%) offer free tiers that cover this class of bot.
Example use case: First-line customer support for an e-commerce store handling 1,000-5,000 user messages per month with AI-generated responses backed by your documentation.
| Component | Typical cost |
|---|---|
| Telegram API | $0 |
| Hosting (small VPS) | $5-15/month |
| LLM costs (GPT-4o mini or Claude Haiku, ~3,000 msgs/mo) | $15-40/month |
| Database (PostgreSQL free tier) | $0 |
| Optional: no-code builder | $10-50/month |
| Total (code yourself) | $20-55/month |
| Total (no-code builder) | $30-105/month |
This is the most common paid-bot archetype in 2026 per public case studies we reviewed. A common pattern is to use GPT-4o mini or Claude Haiku as the primary model (fast, cheap, sufficient for 80% of queries) with occasional fallback to a frontier model for harder questions.
Example use case: News channel, crypto signals bot, or content creator pushing updates to 20,000-100,000 subscribers.
| Component | Typical cost |
|---|---|
| Telegram API | $0 |
| Hosting (mid-tier VPS) | $20-50/month |
| Paid Broadcasts (50,000 recipients, 2 broadcasts/day, 30 days) | $300-900/month |
| Content generation (if AI-assisted) | $20-100/month |
| Database (user list + preferences) | $0-25/month |
| Total | $340-1,075/month |
The Paid Broadcasts line dominates. This is the only archetype where Telegram-side costs actually matter. A channel can minimize costs by spreading broadcasts over longer intervals (8-12 hours) to stay within the free 30 msg/sec rate, at the expense of slower delivery.
Example use case: Telegram-based storefront accepting orders, handling product catalog, and processing payments via Stripe or Telegram Stars.
| Component | Typical cost |
|---|---|
| Telegram API | $0 |
| Hosting (small VPS) | $10-25/month |
| LLM for product Q&A | $10-40/month |
| Database (products + orders + users) | $0-25/month |
| Payment processing | 2.9% of transactions (Stripe), 0% of Stars |
| Optional: no-code builder with Stripe integration | $40-100/month |
| Total (code yourself) | $20-90/month + transaction fees |
| Total (no-code builder) | $60-190/month + transaction fees |
The standout feature here is that Telegram itself takes 0% commission on any transaction, whether physical goods through Stripe or digital goods through Stars. Your only transaction cost is whatever your payment provider charges. This is meaningfully different from platforms like Facebook or Instagram Shops, where platform commissions apply.
Most competitor pages give you "30 messages per second" as the rate limit and move on. This is not the complete picture, and bots that assume it is the only limit get blocked.
The Telegram Bot API has three distinct rate limits that operate simultaneously:
Telegram enforces three concurrent rate limits on bot API messages: 30 messages per second globally across all chats, 1 message per second per individual chat, and 20 messages per minute per group. [2]
Global: 30 messages per second. This is the one everyone quotes. Across all chats combined, your bot can send roughly 30 messages per second. Exceed this and you receive HTTP 429 errors with a retry_after field telling you how long to wait.
Per individual chat: 1 message per second. Even if your global rate is low, sending more than one message per second to the same user triggers throttling. Burst messaging to an individual user, even a short sequence like a welcome flow with four messages fired rapidly, can exceed this limit if you do not add delays.
Per group: 20 messages per minute. For bots in group chats, the limit is much tighter. This is why automated polls or quiz bots in large groups need to space their messages out or risk being rate-limited across the whole chat.
These limits are per-bot, not per-server. Running multiple bots on the same VPS gives you separate limit budgets for each. They are also "soft". Telegram allows brief overages in practice, but repeated violations trigger escalating 429 blocks that can lock your bot out for 30+ seconds at a time.
Paid Broadcasts only raises the global limit
Enabling Paid Broadcasts via @BotFather raises only the global 30/sec limit, up to 1,000 messages/second, at a cost of 0.1 Stars per message above the free threshold. It does not change the per-chat or per-group limits. A bot that hits the per-chat limit because it sent a user four rapid messages in a welcome flow will not be helped by Paid Broadcasts. The fix is to add delays.
The right way to handle limits in production:
Use a library that implements queuing with exponential backoff on 429 errors. grammY, aiogram, python-telegram-bot, and Telegraf all have built-in retry logic or well-documented patterns for it. A common mistake is looping through a user list and firing sendMessage calls as fast as possible. Telegram blocks your entire bot after roughly 30 messages, affecting everyone interacting with it, not just the broadcast recipients.
At this point you have the numbers. The remaining question is which path to take. There are three, and the right one depends mostly on who you are, not what your bot does.
Path 1: Write the code yourself. Cheapest ($5-55/month total) and gives you full control. You pick the language (Python, Node.js, Go, Java, PHP, Ruby, C++ all have mature SDKs), the hosting, and the architecture. Time-to-first-bot is usually 2-8 hours for someone comfortable with basic web development. This path wins when you want portability (the bot runs anywhere), when you need custom logic that does not fit a builder's conversation model, or when you want to keep recurring costs near zero.
Path 2: Use a no-code builder. Fastest (a working bot in 30 minutes to a few hours) and requires no programming. Monthly cost runs $10-100 depending on volume and features. The tradeoff is that you pay a subscription, you are limited to the builder's conversation model, and moving your bot to a different platform later means rebuilding. This path wins when the person building the bot is not a developer, when the business needs to ship immediately, or when the bot is simple enough that a visual flow builder expresses the logic cleanly.
Path 3: Hire a freelance developer. Highest upfront cost (500-5,000 EUR on Upwork or Toptal depending on complexity) but you end up with custom code you own outright. Then you pay only for hosting ongoing. This path wins when the bot needs to do something unusual (tight integration with a proprietary system, specific AI orchestration, complex state machines), when your team has no time to build it, or when you expect the bot to run for years and want to avoid accumulating subscription costs.
A quick decision rule: if you can describe your bot's behavior as a flowchart with fewer than 20 decision points, a no-code builder will work. If the flowchart has 50+ branches or needs to call external services in non-trivial ways, code is faster to write than the builder's logic will be to configure.
The honest picture: not every no-code builder supports Telegram equally. Some have it as a first-class channel with a direct @BotFather token connection. Others support Telegram only through Zapier, Make, or custom code bridges. This section tells you which is which based on live verification of each vendor's current integration, not marketing claims.
Affiliate disclosure: Landbot and MindStudio links are affiliate links. We earn a commission if you subscribe through them at no additional cost to you. Manychat, Botpress, and Voiceflow are not affiliate partners and are included because honest comparison requires it.
Telegram support: Fully native. Connect via Settings → Telegram → bot token from @BotFather. No code, no middleware.
Manychat offers a free tier for up to 1,000 contacts with basic features. Pro plans start from $15/month (or approximately $11/month billed annually) for 500 contacts and scale to $100+/month for 25,000+ contacts. [3]
Strengths: Largest user base in the no-code chatbot space (over 1 million businesses). Multi-channel support includes Instagram, Facebook Messenger, WhatsApp, SMS, email, and Telegram from one dashboard. Deep Shopify integration for e-commerce. "Comment-to-DM" automation is best-in-class for social media funnels.
Weaknesses: Primary focus is Instagram and Facebook, with Telegram treated as a secondary channel. AI features are capable but not as central as MindStudio. Limited to Meta-friendly use cases. If your business is running on TikTok or other non-Meta channels, coverage is partial.
Best for: E-commerce brands and creators who already use Instagram/Facebook and want to add Telegram as an additional channel.
Telegram support: Native. Connects via bot token in the channel configuration.
Botpress offers a free cloud tier with limited messages. Paid cloud plans start at $89/month for more advanced limits. The self-hosted community edition is free, with users paying only for hosting. [4]
Strengths: Developer-friendly visual builder with the ability to drop into code for complex logic. Strong AI agent capabilities with LLM integration. Open-source option for full data control.
Weaknesses: Learning curve is higher than Manychat. Pricing at paid tier jumps steeply. Self-hosted requires ops knowledge.
Best for: Technical teams who want visual development but need the option to extend with code. Open-source requirement.
Telegram support: Native via the channel marketplace. Connects via bot token.
Voiceflow offers a free plan for individual use. Pro starts at $50/month and Team plans start from $150/month. [5]
Strengths: Strong AI agent focus with integrated knowledge base, LLM routing, and conversation testing. Good for complex conversational flows.
Weaknesses: Higher price point than Manychat. Less optimized for broadcast/marketing use cases than for AI assistants.
Best for: Teams building AI-powered assistants where conversational quality matters more than broadcast volume.
Telegram support: Not a native channel, but MindStudio agents can be deployed to Telegram via the platform's integration tools or by connecting agent outputs to a Telegram bot through standard API calls.
MindStudio offers a Free plan with 1 agent, 1,000 runs/month, and 200+ AI models via Service Router with no API keys required. Individual is $20/month plus usage ($16/month billed yearly) for unlimited agents and unlimited runs. Business pricing is custom and includes SSO, audit logs, SLA, self-hosting, and compliance controls. [6]
Strengths: The AI agent focus is real. 200+ models from OpenAI, Anthropic, Google, Meta, and others accessible through one interface with no separate API keys. Build an agent once, deploy to multiple surfaces. Data is explicitly never used for training. Our CTI score for MindStudio is 95/100 (Grade A), the highest in its category.
Weaknesses: Telegram integration requires understanding how to wire MindStudio's agent outputs into a Telegram bot, which is easier than writing everything from scratch but not as point-and-click as Manychat. Smaller integration ecosystem than Landbot or Manychat.
Best for: Teams building AI-first bots where the LLM routing, prompt engineering, and multi-model flexibility matter more than the specific messaging channel.
Telegram support: Not a first-class channel. Landbot connects to Telegram through their APIChat system, which requires either Node.js middleware code (their help docs walk through the setup) or a third-party bridge like Zapier, Make, or Pabbly Connect. This is a real technical step, not a five-minute setup.
Landbot pricing in April 2026: Sandbox (Free) at €0/month for 100 chats and 1 seat, Website channel only. Starter at €32/month billed annually (€40 monthly) for 500 chats, 100 AI chats, 2 seats, Website + Messenger. Pro at €80/month billed annually (€105 monthly) for 2,500 chats, 300 AI chats, 3 seats, Website + Messenger + API channels with advanced integrations. WhatsApp Pro at €160/month billed annually (€200 monthly) adds WhatsApp with 10,000 messages/month included. Business starts from €400/month custom. [7]
Strengths: Landbot's native-feeling visual builder and broad non-Telegram channel support make it strong for multichannel campaigns. If your bot needs to live on Website + WhatsApp + Messenger + Telegram, Landbot lets you build once across four surfaces. Advanced integrations (HubSpot, Salesforce, Stripe) work natively on Pro tier. Our CTI score for Landbot is 85/100 (Grade B).
Weaknesses: Telegram support requires technical work to wire up, which defeats the "no-code" value for Telegram-only bots. WhatsApp features are stronger than Telegram features by a wide margin. Starter plan's 100 AI chats/month is tight for bots using any AI.
Best for: Teams building multichannel conversational experiences where Telegram is one of several channels, not the primary one. Teams that already use Landbot's deep CRM integrations (HubSpot, Salesforce) and want to extend to Telegram.
| Your situation | Recommended builder |
|---|---|
| Non-technical, Telegram-only, need to ship fast | Manychat |
| Technical team, want visual + code extensibility | Botpress |
| AI-agent-first bot, multi-model LLM routing matters | MindStudio (affiliate) |
| Multichannel (Telegram + WhatsApp + Website + Messenger) | Landbot (affiliate) |
| Conversational AI with focus on flow quality | Voiceflow |
| You code and want to avoid subscriptions | Write it yourself with grammY, aiogram, or python-telegram-bot |
We analyzed 12,694 impressions across 167 distinct queries that landed on our Telegram Bots review page between mid-January and mid-April 2026. The breakdown is striking and publishable because, as far as we can find, no other source has this granularity for Telegram-bot-specific searches.
71.6 percent of queries containing 'free'
Across 12,694 impressions on /review/telegram-bots/ between mid-January and mid-April 2026, 49 distinct queries containing "free" or explicit free-pricing language accounted for 3,032 impressions, or 71.6% of total page impressions. General pricing or cost queries accounted for an additional 18.5%. Combined, pricing-related intent represents 90.1% of all search demand to this page.
| Intent category | Queries | Total impressions | % of total |
|---|---|---|---|
| "Is it free" / explicit free pricing | 49 | 3,032 | 71.6% |
| General pricing / cost | 40 | 782 | 18.5% |
| Rate limits | 16 | 181 | 4.3% |
| Other long-tail | 49 | 104 | 2.5% |
| Brand/generic ("telegram bots") | 4 | 69 | 1.6% |
| How to / creation cost | 9 | 68 | 1.6% |
Three things stand out:
Almost all of the demand is pricing confirmation, not exploration. Nine of every ten queries have "free" or "pricing" or "cost" in them. Developers already suspect the API is free. They want confirmation and specifics, often with a 2026 date stamp and ideally pointing to official documentation.
"Official docs" is a frequent query modifier. Specific phrases like "telegram bot api free of charge no fees official docs" (457 impressions alone) show users want primary-source confirmation. Third-party reviews underperform the official Telegram documentation for this intent, which is one reason why our original page was not capturing many clicks despite ranking in the top 10.
Rate limits are a significant secondary intent. 4.3% of impressions are specifically about rate limits, which is a developer-focused technical question that most pricing-oriented content ignores. Our rate limits section above targets this segment specifically.
This data collection is ongoing. We plan to repeat this analysis quarterly across all six verticals in our network, publishing a cross-category search behavior report in Q2 2026.
API pricing and rate limits: Verified against Telegram's official Bot API documentation (core.telegram.org/bots/api) and FAQ (core.telegram.org/bots/faq) on April 19, 2026. Where Telegram's documentation is deliberately vague (the soft nature of rate limits, for example), we note the community-observed behavior and cite its source.
Telegram Stars → USD rates: Calculated using the fixed 200 Stars = 1 TON conversion rate published by Telegram, combined with TON's public market price on the date of publication. The rate table above shows the full range observed since Stars launched in mid-2024.
No-code builder pricing and features: Verified against each vendor's official pricing page on April 19, 2026. Landbot and MindStudio pricing is based on our direct CTI audit of these vendors in the BlockSentient database (CTI scores 85 and 95 respectively). Manychat, Botpress, and Voiceflow pricing is from their public pricing pages.
Query intent data: Aggregated from 12,694 Google Search Console impressions for our /review/telegram-bots/ page between January and April 2026. Queries were classified into intent buckets by keyword pattern matching. Full dataset available on request.
Affiliate relationships: Landbot and MindStudio are affiliate partners. We earn a commission if you subscribe through the links above at no additional cost to you. Manychat, Botpress, Voiceflow, Botract, Chat Data, and freelance developer options are included without any affiliate relationship. Our philosophy, "Documentation, Not Opinions," requires honest comparison even when the recommendation does not pay us.
What we cannot verify and did not claim: We do not have SOC 2 Type II attestation data for any no-code builder except where the vendor publishes it on their site. We do not have real-time TON market data and therefore use published rates as of the publication date; actual exchange rates fluctuate daily. The Paid Broadcasts 10,000 Stars activation threshold is current as of April 2026 and could change with Telegram's policy updates.
Updated: This page is updated quarterly or when any significant pricing changes. Last verification: April 19, 2026.
BlockSentient is part of the Veracity Media Network, a two-person independent SaaS directory operated by Emerging S.R.O. in Brno, Czech Republic. We do not accept paid placements. We audit 349 AI tools using our Compliance Transparency Index (CTI) methodology. Our database shows 94% of conversational AI tools publish public pricing and 90% offer a free trial or free tier, well above the industry average for enterprise software. If you are a tool vendor and believe your information should be updated, contact us.