Langtrace.ai

Open-source observability and evaluations platform designed to improve LLM apps. Get started today for free.

[Jay Thakrar @JayThakrar]

3 AI agents. One lean GTM team. Here's how we scaled Langtrace's go-to-market without scaling headcount.

At an early-stage startup, you don't get to hire your way out of every problem. So we built our way out instead.

Here are the 3 agents that changed how we operated:

→ Agent 1: Lead Enrichment

Every new signup to our managed service fired a Slack notification via webhook.

The agent grabbed that identifier and enriched it across Apollo, ZoomInfo, LinkedIn, and Exa, in seconds.

By the time a prospect finished their first session, I already knew if they were worth pursuing and had reached out on LinkedIn.

Result: 6x improvement in freemium-to-paid conversion.

→ Agent 2: Customer Intelligence

Every week: a full usage summary across our top 50 customers. Expansion signals. Churn flags. All derived from a single key metric — traces generated.
That early warning system paid off in a real way.

It surfaced that one of our highest-usage customers, a $25M-funded platform powering AI tools for 50% of the top 50 insurance brokers, was quietly at risk.

We caught it early. Got in front of them. Captured the integrations they needed. Fed it straight into our roadmap. Retained them.

Without the agent, we would have missed the window entirely.

→ Agent 3: Product Insights

Every customer call, prospects, new signups, veteran users, got synthesized into product gaps and messaging improvements automatically. Objective voice-of-customer research at scale, without a research team.

The common thread across all three: each agent removed one specific bottleneck between signal and action.

GTM teams that build this kind of infrastructure will outpace those that don't.

What's the first process you'd automate if you had the time to build it?

[Langtrace.ai @langtrace_ai]

👀👀

[Karthik Kalyan @karthikkalyan90]

I have long believed this and as a result we have been experimenting on some fresh updates to @langtrace_ai based on our experience building agents with @HeyZestAI.

Processing 100K+ OTLP traces entirely in-browser(zero backend). Some web assembly + parquet magic. Wild what's possible now! Stay tuned!

[Zest @HeyZestAI]

Excited to introduce Zest, a no-code platform to launch AI agents as virtual team mates in Slack.

Zest connects to all your 3rd party apps, including Linear, Notion, Hubspot, Google Workspace, and more. As a result, agents built on Zest are Slack-native, understand your business, and can complete end-to-end tasks across multiple tools.

Any employee can create their own AI agent on Zest; it is built to help your team with everyday tasks, freeing you up to focus on what matters most. Think of it as an extra pair of hands that can help you with things like:

-> Searching in Slack
-> Summarizing Conversations
-> Creating drafts of documents from conversations
-> Converting docs into tickets on Linear

To get started, it’s as simple as 3 steps:

1. Connect Zest to your Slack Workspace
2. Create your AI agents and connect your favorite 3rd party tools
3. Configure and customize your AI agents (optional)

Ready to leverage Zest to augment your team with virtual coworkers? Sign up via link below.

[Karthik Kalyan @karthikkalyan90]

Slack is an incredible product surface for building agents. We have built a pretty powerful platform that lets you create and deploy agents as Slack bots in under 2 minutes. As soon as they are installed, they instantly know everything about what's happening in your Slack that is most relevant to you. They continue to gather this knowledge and are available to assist you at all times. Here's a quick demo of the first time install experience(or the landing experience as @rauchg puts it). Feels truly magical to see it take this current shape after months of hard work. I will continue to share more demos in the coming days. DM me for access if you would like to take this for a spin.

[Karthik Kalyan @karthikkalyan90]

💫New name and new features

DSPyground now features a real time view of the hill climbing GEPA does along with all the prompts it tries as it optimizes. Also, a reorganized UI with light/dark mode toggles.

Real time Prompt E̶n̶g̶i̶n̶e̶e̶r̶i̶n̶g̶Optimization powered by @DSPyOSS GEPA and @aisdk

What else do you need? 👇Jump into the repo and let me know.

https://t.co/uYdeUCXofC

[Karthik Kalyan @karthikkalyan90]

Introducing: @aisdk prompt optimizer for improving Agent trajectories and tool call adherence. Powered by @DSPyOSS GEPA. Fully OSS!

Powered by @aisdk , @DSPyOSS and @shadcn

1. Clone to repo (linked in the last tweet)
2. Run `npm run dev:all` to start it up.

[Karthik Kalyan @karthikkalyan90]

So, the latest hire (powered by Zest) is @vercel agent Slack bot. It can do a few things, but the notable one is,
- Whenever some one mentions about deployment to prod, it proactively jumps in, checks the deploy logs and tags Devin to take a look if there are any errors. If Devin is tagged, the deployment issue is fixed very quickly. All this happens automatically in under 5min.
@vercel agent bot built on Zest which is built using @aisdk and deployed to @vercel = vercelception

[Langtrace.ai @langtrace_ai]

Thanks for the great overview @MervinPraison

[Mervin Praison @MervinPraison]

AI agents are powerful, but you're flying blind without observability.

Langtrace: open-source tool that tracks every API call, cost, and performance metric in your AI apps.

Just 2 lines of code to integrate with @langchain , @llama_index , @crewAIInc .. :

[Mervin Praison @MervinPraison]

AI agents are powerful, but you're flying blind without observability.

Langtrace: open-source tool that tracks every API call, cost, and performance metric in your AI apps.

Just 2 lines of code to integrate with @langchain , @llama_index , @crewAIInc .. :

[Karthik Kalyan @karthikkalyan90]

We created a Slack bot called Zest that has access to all b2b apps we use including Slack itself. Now, building an exec summary got much easier.

Prompt to a live website in <30 seconds 🔥

Powered by @aisdk and v0 platform API.

[Karthik Kalyan @karthikkalyan90]

👀 @SlackHQ is simply the perfect product surface for running agents. Here's quick demo of our slack bot, Zest seamlessly navigating between @NotionHQ and @linear to complete a workflow without ever leaving Slack.

[Jay Thakrar @JayThakrar]

Typical convo with @karthikkalyan90 at 9pm on a Weds...

...building and optimizing the deterministic + nondeterministic logic of an AI agent with enough context engineering to consistently produce high quality outputs.

... what a time to be alive and building. stay tuned for an upcoming and exciting new product launch!

[Karthik Kalyan @karthikkalyan90]

💯 We have built an agent called Zest that runs purely on Slack. It has access to all b2b tools we use and can run point on gathering everything you need to complete workflows for you. The goal here is not to do everything in the background but, to give the user essential control and information to complete the last mile of the work in much less time without losing focus or context switching. This has been a huge boost in productivity for us.

Here's a video of Zest gathering the details of the latest ticket from Linear and then the user(me) assigning the task over to Cursor agent which completes and creates a PR.

[Langtrace.ai @langtrace_ai]

📢📢Enhancing AI Observability with @MongoDB and Langtrace
https://t.co/miCxpESwDT

[Langtrace.ai @langtrace_ai]

@coinbase mafia 🦾

[Coinbase Ventures 🛡️ @cbventures]

Tonight we are celebrating the 100+ @coinbase alum-turned-founders pushing crypto forward.

Help us kick it off: What's something you learned during your time at CB that led you to where you are today? 👇

[Ashpreet Bedi @ashpreetbedi]

Langtrace: https://t.co/WHttYnKOKR

@langtrace_ai

[Langtrace.ai @langtrace_ai]

Exciting news for graph enthusiasts! Langtrace now seamlessly integrates with @neo4j and @neo4j GraphRAG! 🚀

Unlock the power of graph databases for your language model applications and gain unparalleled context and reasoning capabilities. With Langtrace, you can now trace and debug your LangChain applications leveraging Neo4j and GraphRAG, providing deeper insights into your retrieval augmented generation workflows.

Ready to explore the synergy of LLMs and graph data?

Discover more here: https://t.co/c1KDD94K0G

[Langtrace.ai @langtrace_ai]

Thrilled to share some exciting news! Langtrace is now officially recognized as an External Trace Processor by @OpenAI! 🎉

This integration allows developers to seamlessly leverage Langtrace's powerful tracing and debugging capabilities directly within their OpenAI workflows. Gain deeper insights into your language model applications, identify bottlenecks, and optimize performance like never before.

Ready to experience enhanced observability for your OpenAI projects?

Learn more here: https://t.co/wzwwbIYInb

[Langtrace.ai @langtrace_ai]

📢 📢 If you are building multi agent systems in 2025 and dealing with the challenge of observing them at scale, do RSVP to this session by our cofounder, @karthikkalyan90 and Guangya Liu from IBM at #KubeCon 2025.

@opentelemetry has emerged as a powerful framework for observability in cloud-native applications, but how does it apply to the intricate needs of AI Agent observability? This session explores the journey of leveraging OpenTelemetry to monitor, trace, and analyze AI Agents. We’ll cover key challenges such as capturing metrics for multi-agent systems, tracing inference workflows, and correlating AI-specific data like model performance and decision latency.

https://t.co/U49jvMYkcj

[Langtrace.ai @langtrace_ai]

- https://t.co/q4sjQy0fZP
- https://t.co/tnl61kAgqJ

[Langtrace.ai @langtrace_ai]

🚢 We are excited to announce that Langtrace now supports OpenTelemetry based tracing for @OpenAI 's new Responses API, featuring full OpenTelemetry (OTEL) compatibility. This integration enhances your ability to monitor and trace AI-driven applications, providing deeper insights and greater flexibility.

Why This Matters
⏩ OpenTelemetry (OTEL) Compatibility: With OTEL support, Langtrace ensures that traces are standardized, allowing for seamless integration with various observability platforms. This standardization facilitates richer metadata capture, aiding in debugging and performance optimization.
⏩ Simplified Integration: Incorporating Langtrace into your OpenAI Responses API setup is straightforward. With minimal configuration, you can begin collecting valuable trace data without extensive modifications to your existing codebase.
⏩ Flexibility and Vendor Independence: Langtrace's adherence to OTEL standards means you're not confined to a single observability vendor. You have the freedom to route your trace data to any OTEL-compatible backend, such as @elastic APM, @grafana Labs, @datadoghq , and more, based on your preferences and requirements.