I use ChatGPT and Claude for maybe half my research now. Copy a link, open new tab, paste. I realized I do this constantly – and none of it shows up in analytics.
So I built Loamly. Two parts:
1. Shows what AI platforms say about your brand. Real queries, real answers across ChatGPT, Claude, Perplexity.
2. Detects when AI-referred visitors land on your site. Uses RFC 9421 signatures for ChatGPT Operator, navigation timing for the rest.
Free: 5K credits/mo. Paid: only if you have more than 50k views/events a month. Starts at 89 usd/month
The repo looks solid; wondering where Part 1 (synthetic prompting) lives though? Or are you suggesting that seeing traffic from a bot helps you learn of the questions LLM client users asked?
1. Brand monitoring — Loamly runs queries against ChatGPT, Claude, Perplexity etc. and shows you what they say about your brand. That part runs on our side (needs API keys, scheduled jobs), so it's not in the open source repo.
2. Traffic detection — this is what's in the repo. Detects when AI-referred visitors hit your site using HTTP signatures and timing patterns.
They work together but you can use either on its own. Part 1 shows what AI is saying, Part 2 shows when it actually sends you traffic.
The idea is to give you a full picture: synthetic prompting shows how AI recommends you, traffic detection shows when that turns into real visits. Connecting both lets you actually measure the impact — and the RAG chat (Ask Loamly) can tie it all together.
Let me know if anything's unclear — happy to dig into the details.
I use ChatGPT and Claude for maybe half my research now. Copy a link, open new tab, paste. I realized I do this constantly – and none of it shows up in analytics.
So I built Loamly. Two parts:
1. Shows what AI platforms say about your brand. Real queries, real answers across ChatGPT, Claude, Perplexity.
2. Detects when AI-referred visitors land on your site. Uses RFC 9421 signatures for ChatGPT Operator, navigation timing for the rest.
Free: 5K credits/mo. Paid: only if you have more than 50k views/events a month. Starts at 89 usd/month
If something's broken or confusing, tell me.
The repo looks solid; wondering where Part 1 (synthetic prompting) lives though? Or are you suggesting that seeing traffic from a bot helps you learn of the questions LLM client users asked?
Good question!
So there are two parts:
1. Brand monitoring — Loamly runs queries against ChatGPT, Claude, Perplexity etc. and shows you what they say about your brand. That part runs on our side (needs API keys, scheduled jobs), so it's not in the open source repo.
2. Traffic detection — this is what's in the repo. Detects when AI-referred visitors hit your site using HTTP signatures and timing patterns.
They work together but you can use either on its own. Part 1 shows what AI is saying, Part 2 shows when it actually sends you traffic.
The idea is to give you a full picture: synthetic prompting shows how AI recommends you, traffic detection shows when that turns into real visits. Connecting both lets you actually measure the impact — and the RAG chat (Ask Loamly) can tie it all together.
Let me know if anything's unclear — happy to dig into the details.