Hey there,
I spent the morning today looking at the ecosystem of tokens built around Telegram trading bots. Cumulatively, they account for $147 million in market capitalisation. The largest amongst them, Unibot, has a capitalisation of $54 million as I write this. The tokens linked with these products have undergone boom and bust cycles.
Unibot, the largest among them, is down 75% since mid-August. A more recent entrant named BANANA is still undergoing price discovery. This piece will not bore you with the specifics of the activity on these Telegram-based apps. I highly recommend referring to this Dune dashboard to track activity.
Instead, I will summarise what they do and why market participants run towards them today. We'll then explore what kind of applications could be built on them.
The Present

Telegram's current bot landscape (in relation to crypto) is a spectrum. On one end, you have the usual bots that bring in news about Web3 events from platforms like Twitter (now X) or RSS feeds. They are not crypto-native products but use Telegram for distribution. Wu Blockchain, DeFiLLama and The Block, each have Telegram news channels that relay updates.
Tools that use on-chain events and give context alongside them are usually handled by smaller startups or creators For instance, Jobstash has a feed of new jobs listed by Web3 native platforms. Officer CIA has a feed with a curated subsection of security related articles.
The next layer of products are wrappers on AI tools like ChatGPT. This mix of products allows users to train a bot (on an external website) and introduce it to a Telegram community or use Telegram as an interface to communicate with a bot. PaalAI, for instance, allows you to ask questions like 'What is today's top-performing coin?'
Similarly, NoiseGPT and ChainGPT allow users to produce deepfakes using Telegram as an interface. While elements of crypto can be pretty active in these products, many of them are not 'Telegram applications' but interfaces. (I'll explain the difference between the two shortly.)
Tools like Collab.Land and Guild.xyz are verification tools for token- or NFT-gating communities on Telegram. Here's how they function: a user goes to these products, links their wallet, and signs a message (from their wallet) to prove they own the wallet. The product then checks if the wallet being linked has the necessary number of tokens or NFTs to permit access for a user. In such a product, the "on-chain” function is verifying asset-ownership.
The next group of apps on the spectrum have been around since at least 2019. Santiment, for instance, tracks on-chain events and notifies users. Similarly, Nansen tracks the movement of assets from funds and informs users through the app. While Santiment uses raw data (X wallet moved Y tokens to exchanges), Nansen gives additional context (the owner's name of wallet X) when notifying users.
These are not Telegram applications either. They notify users of the movement of assets through Telegram as a medium.
The tokens that have been all the rage on Twitter in the past few months are full-fledged applications. The difference between an application and an interface (in my definition) is that an application can function fully on Telegram and leave an on-chain imprint without requiring external wrappers (like ChatGPT).
Both Unibot and Banana Gun permit users to set up new wallets, track the launch of new meme tokens and allocate a portion of their wallet balances towards buying these assets without signing transactions each time.
These applications are unique because they create verifiable imprints on-chain. Products like LootBot take this one step further by offering a SaaS-like option for users to farm airdrops on-chain. I could not test the product personally, but according to their Telegram handle, they have enabled some $4.5 million worth of transactions on-chain for users looking to game airdrops.
One of the businesses that have been making waves is being built by a community member at Decentralised.co. Cielo has Telegram native wallet trackers that can track thematic wallet lists or tokens. The interface (on the website) allows users to select a pre-curated list of wallets such as the top 50 earners on FriendTech or wallets that have made millions trading meme-coins in the past.
Users get a notification anytime a user from these wallet subsets makes a transaction. Part of what makes Cielo interesting for me is the number of chains and early-stage types they are quick to support.
For instance, quite recently, they were one of the earliest to allow users to track FriendTech key’s trades, historical P/L and user behavior through Telegram. There is a complex engine parsing data in the background. But the medium of interaction (chats on Telegram) fuels the product's growth.
According to the founder (Matt Aaron) , using Telegram as a medium for distribution has unlocked niches they previously didn’t know about. For instance, a subset of their users have been using Cielo to track NFT loans on Blur. (Presumably to buy liquidated NFTs at a discount). These users receive a notification via Telegram each time a loan is made or NFTs get close to liquidation.
The team behind Cielo began noticing how multiple Telegram apps can be combined to create easier user flows. For instance, forwarding a message from Cielo to Maestro (a trading bot) automatically imports the smart contract addresses of the tokens involved. This allows users to place orders at a fraction of the time it usually takes to do the same on Metamask and Uniswap.
Matt told me (while simultaneously tending to his 1-year-old) that timing becomes of the essence when it comes to Telegram-based applications. Being able to roll out event-specific updates within meaningful periods helps with growth. For context, during the BALD token hype, Cielo sent some 2.5 million notifications on Base (the chain) alone in 48 hours. They track over 20 different blockchains. That is scale, facilitated by Telegram. They have over 44,000 users as of writing this.
All of these are interesting, but what is the opportunity set? To explore that, one should understand what Telegram enables today.
The Assets
At its crux, Telegram is a distribution medium. If you treat it like a social network (like Facebook in 2008), it will become easier to understand why applications have been growing virally. What it enables is surpassing the app store's restrictions to create chat-based interfaces that function off a wallet generated in the product.
Naturally, security trade-offs exist, as the user often does not entirely control the wallet. But for smaller transactions that do not require a cold storage wallet, Telegram enables interacting with a product without switching applications.
In this regard, it is similar to WeChat, except that its APIs and policies are sufficiently open to allow developers to create applications on top of it. Telegram has a wallet feature natively enabled on it. It allows users to hold assets like Bitcoin and several stablecoins.
The purpose of this integration is the possible monetisation of large chats like LobsterDAO. For instance, the community on Telegram has 23,000 active members. It is reasonable to think 10% of that community would convert to paying $20 a month, which should be $40k+ in revenue for the person running it.
As a platform, Telegram is inching closer towards integrating crypto natively into its product. This allows developers to target crypto-native users directly in the chat application. But what if a person does not already hold crypto? Telegram has a solution for that, too.
They just rolled out P2P markets and credit card purchases of crypto in the product. So, an application can launch and collect payments (in stablecoins) from credit card users worldwide in a few clicks. You could not historically do this on other social networks like Twitter (now X), WeChat or Meta if you were a crypto application.
We have a distribution medium, on-ramps and a native wallet. We also have multiple communities with tens of thousands of people. This is similar to when Zynga relied on Facebook for its growth in the late 2000s.
I often wonder if Zynga grew due to Facebook or if Facebook grew due to Zynga. One way to think of it is like this: As Facebook expanded beyond the English-speaking world, many users in emerging markets did not have experience with the mannerisms that came with interacting online.
Games like Farmville made the internet more usable at a time when much of the emerging world was slowly coming online. You no longer had to write a 500-word punchy status update for Facebook to be relevant to you as a user. Even the user who wanted to play Farmville (like me in 8th grade) had something to do for our daily dose of dopamine. Why does this matter?
Chat-based applications on Telegram could make crypto more relatable over the next few years. Not everybody wants to be the owner of a limited-edition monkey face. The market where participants put tokens in smart contracts and get a yield on it saturates rapidly in a bear market.
We expand the user base only when social applications that allow retail participants to engage without layers of complexity involved. For instance, AR seeped into the public psyche with PokemonGO in 2016. More recently, PlayStation's VR headset and Apple’s Vision Pro have brought virtual reality to the masses. Technologies scale when they become accessible to individuals. Telegram, as an interface, could catalyse this for crypto.
But why do I say so? We must look at what Telegram could enable and the mix of applications that could be built on it to understand.
The Possibilities
Long before FriendTech became all the rage on Twitter (now X), the Collab.Land bot allowed token and NFT holders to enter gated chats and communicate. At its core, a person's FriendTech key has the same value proposition as an NFT (for entering a permitted chat). It gives access to a creator (or brand) with whom you can chat. The difference is that FriendTech released a stand-alone app with keys on a bonded curve. Where NFTs had flat prices, bonded curves surged substantially in value.
FriendTech also tapped into Twitter users' social graphs from Twitter to enable discovery and used a points-based system to incentivise users to buy keys. But all of this could have also been replicated on Telegram. Nothing stops Telegram (or a third-party developer) from releasing a FriendTech clone that allows all of the user's contacts to discover, speculate and interact through the chat application.
The core barrier here is that people already talk to one another on Telegram without holding keys. If I suddenly decide I will only respond to Siddharth if he owns a key of mine on Telegram, this would lead to some unpleasant interactions. Especially given that Siddharth (like everyone else on the chat app) is habituated. to getting responses from me on Telegram without holding a speculatory key.
The opportunity subset is for large community owners looking to monetise. Telegram has multiple communities with tens of thousands of users, yet no monetisation model. If these communities decided to monetise with a key-like system (as on FriendTech), creators could monetise at a massive scale with minimal friction. If a royalty model like the one on FriendTech is enabled on Telegram, it could create sufficient revenue to enable a DAO, which could be responsible for curating and scaling a community.
FriendTech on Telegram in itself is not massively interesting. What excites me is the interface between a mobile application and on-chain activity becoming far simpler. For instance, a bot could generate an NFT by simply uploading a photo. Similarly, a user's location data could be captured from their Telegram interface. Imagine going to a concert, opening Telegram, uploading your location to prove your presence and getting a NFT minted directly to your wallet.
Each of these, on their own, is not groundbreaking technology. But they make the possibility of blurring lines between on-chain and on-device far easier.
For instance, an NFT could be issued whenever someone verifies they visited a cafe. Then, the number of NFTs a person owns could unlock discounts or be traded for dollar amounts in a free market built directly on Telegram. Much like how eBay "normalised" auctions on the internet, Telegram as an interface could normalise conducting on-chain actions without ever going through MetaMask.
One way this could play out is by tracking a third-party user's on-chain transactions (like a fund's) and copy-trading the same through a chat-based interface. In such a model, a bot would track on-chain events, notify a user, and ask if they'd like to replicate the same.
A user could then change variables (the amount, gas costs, etc.) and initiate the transaction through the bot. (The example I mentioned above involving Cielo and Maestro bot does this function across two apps. A smart developer could bring this into a single interface and chart a small fee).
Alternatively, products could emerge to track a user's on-chain behaviours and offer alternatives. For instance, if I had a loan on Aave and could get a lower interest rate through Compound, a Telegram-based bot could notify me of the arbitrage and initiate a transaction on my behalf. In all of these cases, Telegram is used as an interface.
At its core, Telegram applications that scale would follow a simple pattern: enabling high-frequency transactions at an extremely low cost. The immediate use case for such an application is (ironically) gambling. Telegram allows users to roll the dice or run a slot machine through chat-based interfaces today.
What truly interests me amongst these applications is the (probability of) training chatbots in the learning of an entire community. I keep going back to them as an example, but consider that LobsterDAO has 4 years of DeFi-power users talking to one another in their community.
It has a knowledge base of 23,000 users sharing links, stories, fears and hopes in the chat. What if an AI could summarise the knowledge of the community? In such a model, the ownership of a token or NFT could (theoretically) gate access to an AI that could tap into the knowledge base of a community's expertise to provide user insights.
This may seem far-fetched, but we have already seen requests to create such a bot within our community at Decentralised.co. We did not accept it because open communities do not (generally) own the conversations generated by their users. Collecting all the members' consent would be a daunting task in itself. But it is quite possible, that in the future, we see AI Bots trained on a community’s interactions. These bots could then be paywalled and used by members who don’t want to search the community’s chat logs.
As an interface, Telegram facilitates an opportunity subset (of on-ramps and interface) among a concentrated user base that understands crypto well. The applications built on it could go viral due to the relative ease of discovery and on-boarding. Developers building social apps, on-chain games and AI-based products stand to benefit the most from building on Telegram as a distribution mechanism. Does Telegram have incentives to kill such applications if they scale? Absolutely.
History offers some precedence here. In 2010, Zynga was under threat of being shut down by Facebook if they chose not to use Facebook’s Credit system. It was a payment method on Facebook that took 30% of the revenue as Apple does in its stores today. So, anytime somebody made a virtual purchase on Farmville, Facebook would make a cut of that transaction. As of 2012, nearly 19% of Facebook’s revenue came from such in-game purchases.
The partnership was to end in 2015 when Zynga could break out and launch its games as a stand-alone firm on its own website without paying Facebook any royalty. Did that transition eventually happen? Not really. By 2016, the building Zynga operated out of was worth more than the firm itself. This fall from grace shows what happens when a distribution medium gets cut off.
Platforms like the App Store can command their platform fees because of the distribution they enable. Today, Telegram does not charge a royalty fee. But if they see large troves of users interacting with crypto-native products, it is not far-fetched to believe they have an incentive to enable fees. Does that make it a bad place to build apps?
Quite frankly, no. You have a product where crypto-native users spend most of their time with an integrated wallet. Building products with Telegram as a distribution mechanism is one way to keep the CAC at an absolute low when resources are scarce. Build, where the users are.
Thinking of shawarma,
Joel John
Disclosure: I am a moderator on LobsterDAO’s Telegram community.
Distribution with Ease of transacting ( Minimal friction) is key. One of the things is that Telegram API and Data model for bots is pretty good and there are good libraries in python. Much can be done without directly interfacing with the MTproto layer.
Analogy of Tensorflow being much more friction then Pytotch in ML and esp LLM development
Thank you so much for mentioning my channel!