Straight answers. No marketing. If your question isn't here, the troubleshooter covers specific failures, and the developer docs cover how everything works under the hood. these answers reflect the 0.2.0-alpha.1 release.
No. The bot never holds your keys or your funds. Every tip is signed by the tipper in their own wallet, and the money moves directly from their wallet to yours on-chain.
The bot's job is just to notice the payment happened and post a thank-you in your chat. It can't spend your money, can't freeze anyone's funds, can't rug. The worst thing a broken bot can do is fail to announce a tip — the tip itself still lands.
No, but you need to be comfortable copy-pasting things into dashboards. The configurator walks you through setup and generates a bundle you upload to whatever host you pick.
If you've ever deployed a personal website on Netlify or Vercel, you can run this. If you've never touched a terminal, Railway is the friendliest path — about 10 minutes, no command line required once the CLI login is done.
Comfortable with a terminal? Self-hosting on your laptop or VPS is cheaper and gives you full control.
No. We take no fee, ever. The project is free and source-available.
The only costs you'll see:
If something costs you more than that, something's wrong. Open the troubleshooter.
Four chains: XRPL, Base, Arbitrum, and Solana. Enable any combination. One, all four, anything in between.
One deployment watches all enabled chains at once. You don't run four bots — a single Node process opens connections to every chain you configured and listens to them in parallel. Same for chat platforms: configure any combination of Discord, Telegram, Twitch, and X, and the same running bot serves all of them. Configure Base, Arbitrum, Solana, Discord, Telegram, and Twitch, and you have one process doing all six jobs with no database, no coordination layer, and no duplication.
You don't pick a chain for your audience. Your audience picks for themselves — the bot shows tippers a signing link for whichever chain they choose in their wallet. Creators who enable more chains catch more tips.
Each enabled chain needs an address you control. The configurator validates the address format before generating your bundle.
Note on hosting: you deploy to one hosting provider (Railway, Render, Fly.io, or self-host) — not all four. That one deployment is what runs the multi-chain, multi-platform bot.
Default configuration:
XRPL supports USDC and RLUSD too, but those require you to set up trust lines in your XRPL wallet first — a one-time signing step that costs ~2 XRP reserve per asset. We default to XRP-only to skip that. See the wallet guide for the trust-line walkthrough if you want stablecoins on XRPL specifically.
None that persists. No database. No analytics. No accounts. No telemetry. No login.
The only state the bot keeps is a 60-second in-memory buffer of recent chat messages, used for pairing EVM tips to the message that preceded them. It holds only the message text — not usernames, not user IDs, not anything identifying — and forgets each entry after 60 seconds. This is needed because EVM chains don't support on-chain memos. XRPL and Solana use native memos instead, so they don't use the buffer at all.
The configurator runs entirely in your browser. Your tokens and addresses never leave your machine. When you click Generate, JavaScript in your browser zips up your bundle locally. Nothing is sent to us, ever.
Audit it. The bot is small — about 3000 lines across 24 files — genuinely readable in an afternoon, or 30 minutes for the parts that matter.
Three honest paths:
package.json against npm.src/api/server.ts, then each chain adapter. You'll understand the bot.AI review tools are explicitly welcome. Paste the bot into Claude, ChatGPT, Copilot, Cursor, or any code-review assistant and ask "is anything suspicious here." Most will give you a useful answer in minutes.
Verify the SHA-256 hash of your download matches the published value before running. See the evaluate guide for more.
Yes. 0.1.0-alpha.7 shipped the MVP with Discord and Telegram. 0.2.0-alpha.1 added Twitch and X — completing the originally-scoped four-platform surface. As of 0.2.0-alpha.1, the canonical release is feature-complete relative to that original design.
When CVEs emerge in dependencies — which they will, that's the npm ecosystem — we cut a patched version, update the hash, and publish the new tarball. Patches are announced on the Releases page. No email list, no push notifications, no RSS yet — check that page periodically or bookmark it. If you deployed over 30 days ago, run npm audit in your bot folder as a sanity check.
We don't intend to add new chains or platforms to the canonical release. The four-chain, four-platform surface is what this project is. If you want a fifth chain or a different platform, forks are explicitly welcome under BSL 1.1 — the small surface area and chain/platform-agnostic adapter pattern make that practical.
Nothing. To you.
Your bot runs on your infrastructure, with your keys, watching your addresses. If this site disappears tomorrow, your bot keeps running. Tips keep landing. The source tarball is already on your host. You don't need us.
What you would lose is a place to download new versions and find updated docs. Mitigate by saving the SHA-verified tarball locally — that copy keeps working forever, on whatever hardware you decide to run it on.
That's the whole point of non-custodial. Nobody is in the loop between your audience and your wallet.
Not officially. The source ships as a tarball at justthetips.dev/downloads. This is a deliberate choice — it keeps one canonical copy with a verifiable hash, and doesn't tie the project's existence to any particular platform.
The license (BSL 1.1, converting to Apache 2.0 in four years) allows fork and modification. You're welcome to put your copy on GitHub, run a fork, or mirror the source anywhere you like.
If a GitHub repo becomes the right thing later, we'll make one. For now the tarball plus hash is the source of truth.
If your question isn't here, try one of these: