The experiment: people submit their crypto loss stories through a form. The agent validates the text with GPT, checks for duplicate wallets, and then sends a small token airdrop on-chain as a symbolic acknowledgment. Every action (send, skip, duplicate) is logged in a public ledger.
Around that, I added supporting agents:
Drop Announcer — posts confirmed drops with proof links to X/Twitter + Telegram.
Bug Fixer — watches for stuck rows and resets them safely via Telegram approval.
Shadowline — scans X mentions and (rarely) replies/likes/retweets when GPT approves the context.
So far the system has processed 250+ submissions and 200+ on-chain drops without manual intervention. Everything is orchestrated in n8n, with GPT handling validation and copywriting, and JSON logs providing transparency.
Repo (flows + docs): https://github.com/GlitchWriter/txn0-agent-flows
Project site (submission form): https://txn0.io
This isn’t financial advice or a token promotion. It’s a case study in how far no-code + AI orchestration can go, even coming from a non-developer background. I’d love feedback, especially from others building complex automations with AI.