Most underrated data source in a company: your AI agent's conversations. Your agent talks to more customers than any employee, but the data it generates goes nowhere. Latitude.so changes that, see how:

Current observability tools don’t capture when users have unmet expectations, are frustrated, or about to churn. @trylatitude shows you how people actually use your agent and alerts you when it silently fails. Open source and MIT licensed. Try it now:

1. Observe: Observability built for agents (not just LLMs): sessions, latency, cost, and automatic tagging for unusual behavior + specific dashboards for tools and users.

2. Understand: See how people actually use your agent: what they ask for, how they feel about it, and signals like satisfaction, frustration, escalation, and churn.

3. Fix: Get alerted in Slack if your agent fails or behaves unexpectedly, and fix it directly from your coding agent via MCP. Create and upload datasets to run regression testing before shipping.

@trylatitude Imagine paying for user interviews while your AI agent has already had 100,000 conversations. 👀

@trylatitude Today this is the third company I have seen that collects content from AI in some way. Enough to call it a trend. This will help developers understand the systems agents are building. Thanks!

@trylatitude I don’t like the word “observability”. I mean, I get that everyone uses it, but I don’t like that it implies we have to “observe” things in order to find out what’s happening. I like you guys are focusing on providing insights, and not only giving people ways to “observe”.

@trylatitude that's exactly right, we prefer "conversation intelligence" but most people don't know what that is (yet)

🚀🚀🚀

@trylatitude Gotta try it our with my agents

@trylatitude looking forward to your feedback 🙏

@trylatitude haha yes, opening some issues on gh next

@trylatitude Congrats @heycesr and team

@trylatitude appreciate the support 🙏

@trylatitude This is an absolute gem of an insight! 💎 The idea that AI agent conversations are the most underrated data source is spot on.

@trylatitude many hidden insights in those conversations

@trylatitude actually shipped a genuinely solid product in the agentic space.

@trylatitude thanks man!

@trylatitude Gooo! Nice one!! Upvoted

@trylatitude can't wait to try this with my ai agents

@trylatitude let us know how it goes!

@trylatitude that’s a real problem to be solved, great launch team

@trylatitude appreciate it marcus. yeah it's been the most requested thing from teams running agents in prod, felt good to finally ship it.

@trylatitude Companies are sitting on a goldmine of insights and barely using it.

@trylatitude The most valuable customer feedback today is hiding inside AI conversations.

@trylatitude thanks for making it open-source as well! i’ve been vibecoding something similar internally for my app, gonna give Latitude a try!

@trylatitude Congrats on the launch 🎉, this looks super useful

@trylatitude thanks dude!

@trylatitude "Observability" was always the wrong word for AI agents.

@trylatitude right. observability shows you the trace. agents need something that finds the issue and helps you fix it. watching isn't enough anymore.

@trylatitude Great job once again, @heycesr and @trylatitude! 👏👏👏 So useful. .. and blown away that you're releasing this as open source! Let's catch up :)

@trylatitude Your agent is basically running user interviews 24/7. The fact that most teams don't analyze those conversations is kind of crazy.

@trylatitude imo, the semantic search alone is worth the install. curious how well the clustering handles messy, half-broken user requests.

@trylatitude This is gem for sure!

@trylatitude biased but agree

@trylatitude agent conversations are free user research

@trylatitude 🎯

@trylatitude The semantic search feature alone is a huge win. Being able to ask "where users got frustrated" and instantly find the conversations is incredibly useful.

@trylatitude Very cool stuff. Congrats on the launch

@trylatitude thanks pedro!

@trylatitude Very cool to see focus on agent observability tools. If anyone is shipping agents to real users, this is worth a look. Open source & MIT licensed!

@trylatitude The conversation data generated by agents is an untapped goldmine for product, support, and retention teams.

@trylatitude The MCP integration is what stood out to me. Finding issues is one thing, fixing and validating them directly from your coding agent is where the real value is. 🚀

@trylatitude Never thought of agent conversations as a data source tbh

@trylatitude soon it'll be one of the most important ones

@trylatitude Regression testing + MCP-based fixes sounds like a powerful feedback loop for shipping reliable agents

@trylatitude real claymation or ai ooc?

@trylatitude Love that this focuses on signals instead of dashboards. I don't want 1,000 traces, I want to know what's actually broken.

@trylatitude Cool stuff Gotta try this out

@trylatitude this is wild, excited to try this out

@trylatitude let us know!

@trylatitude Love the focus on understanding why users struggle, not just tracking metrics.

@trylatitude appreciate it. metrics just tell you something happened, not what to fix.

@trylatitude So glad to finally see an open source tool to monitor my agents!