TraceRoot.AI (YC S25)

@TracerootAI now has a span timeline view 🚀 Every span as a Gantt-style bar — latency hotspots jump out instantly. Click a bar → jumps to the span in the tree. Works on live traces too, bars stream in as the agent runs!

Kimi K2.6 is now in TraceRoot! PR: github.com/traceroot-ai/t… Thanks @lqiu03!

Claude Ops 4.7 now works out of the box in @TracerootAI (just dropped 👀) → debug your agents → trace failures → auto-suggest fixes Big thanks to contributors 🙏 PR: github.com/traceroot-ai/t…

@TracerootAI just got a lot easier to integrate — no manual instrumentation required. 🔧 TraceRoot Skills — works with your coding tools Auto-instruments your codebase from inside your coding assistant. Codex, Claude Code, Antigravity — all supported. One command: `npx skills add traceroot-ai/traceroot-skills` ⚡ Skills are now part of onboarding Sign up → generate API key → run a skill → send your first trace → connect Github. Under 2 min. Free tier, no credit card. When something breaks in your agents, our agent traces the failure and fixes it. Give it a try 👇

RT @xinwei_97: 🚀 TraceRoot.AI (YC S25) v0.1.2 - now with Claude Agent SDK, OpenAI Agent SDK & Gemini support! Full observabilit…

Most AI debugging tools lock you into their models. TraceRoot does the opposite. → BYOK (Gemini, xAI, DeepSeek, Kimi, GLM, …) → Mix & switch providers anytime → Your keys never leave your env Powered by pi-mono - same harness behind @openclaw No lock-in. No black boxes.

Your AI agent runs but it's slow, expensive, or just… off. TraceRoot gives you full-stack tracing to pinpoint exactly why — linking latency spikes, cost blowups & quality regressions directly to the commit that caused them. Want to optimize your AI agents? Let's talk!

Your AI agent fails in production. Now what? TraceRoot tags every trace with your git version, then our AI debugger spins up a sandbox, connects to your GitHub, and pinpoints the root cause across commits, PRs, and issues. Watch it in action ↓

Been quiet for 3 months. Here's what I was building 🧵

@TracerootAI v0.1.5 is out! 🚀 Multi-trace selection: View and compare multiple traces side-by-side with simple checkboxes. Redesigned UI: A cleaner, more consistent experience across traces, logs, and the AI agent chat. Faster agent responses: Significant speedups across the board. Check it out!

@TracerootAI is making something people want. Big thanks to @boltohq for this incredible gift! #YCombinator

My 5 Favorite Launches from Product Hunt – Aug 25, 2025 - Trace @trace_so, @timcherkasov - Qoder @qoder_ai_ide, @Sandy200wh - Onlook for Web @onlookdev, @D_R_Farrell - TraceRoot AI @TracerootAI, @zechengzh, @xinwei_97 - AI Elements by Vercel @vercel, @rauchg, @haydenbleasel

Please help us to get top 3 on Product Hunt🙏 producthunt.com/products/trace…

We are live on Product Hunt! producthunt.com/products/trace… We built TraceRoot because modern observability tools are siloed and reactive. Debugging shouldn’t feel like detective work across logs, traces, metrics, code, and endless Slack threads. TraceRoot is an AI-native, open-source debugging and observability tool connects logs, traces, metrics, code, and team discussions. Please help us to get top 3 on Product Hunt🙏 #YCombinator #ProductHunt #AIAgents

Today is the day... We're live on Product Hunt!! Dropping by and saying hi in the comments would mean a lot 🥰

Production bugs piling up? We’ve got something for you. Stay tuned 🔧

First release for @TracerootAI (YC S25) v0.1.0 is out! 🚀 You can now self-host TraceRoot platform with one terminal command. Just run a single Docker install script and you’re up and running. Our UI is completely migrated to shadcn-based, way sharper. Good progress on BYOM (Bring Your Own Model) support. Now you can plug in your own API keys for OpenAI or Groq. Big shoutout to new contributors who jumped in to help!

🤯🤯🤯 GPT-5 is Truly a Flagship Model! Here we benchmarked 𝗴𝗽𝘁-𝟱 on its real-world debugging capability comparing against 𝗼𝟰-𝗺𝗶𝗻𝗶, 𝗴𝗽𝘁𝟰-𝗼, and 𝗴𝗽𝘁-𝗼𝘀𝘀-𝟭𝟮𝟬𝗯 on the @TracerootAI platform. 𝗦𝗲𝘁𝘂𝗽: The models tried to resolve 10 real world bugs requiring: 1️⃣ Identifying the bug from traces and logs 2️⃣ Writing the code fix 3️⃣ Using GitHub tools to create a PR 𝗥𝗲𝘀𝘂𝗹𝘁𝘀: 📊 𝗴𝗽𝘁𝟱: 90% success 🤯, 0% PR tool fails 🤯, 10% PRs didn’t fix the bug. 📊 𝗴𝗽𝘁𝟰-𝗼: 0% success, 30% PR tool fails, 70% PRs didn’t fix the bug. 📊 𝗼𝟰-𝗺𝗶𝗻𝗶: 50% success, 20% PR tool fails, 30% PRs didn’t fix the bug. 📊 𝗴𝗽𝘁-𝗼𝘀𝘀-𝟭𝟮𝟬𝗯: 20% success, 10% PR tool fails, 70% PRs didn’t fix the bug. 𝗖𝗼𝗻𝗰𝗹𝘂𝘀𝗶𝗼𝗻𝘀: ✅ 𝗴𝗽𝘁-𝟱 is insane 🤯 It almost correctly fixes all @TracerootAI 's selected real-world bugs ✅ 𝗴𝗽𝘁-𝟱's solution is more general than o4-mini etc. ❌ 𝗴𝗽𝘁-𝟱 "thinking" takes more time and makes the whole debugging process around 3x slower (compared to o4-mini) on @TracerootAI 's platform. Also, GPT-5 is available on @TracerootAI 's platform. Feel free to register and try to debug with GPT-5 auth.traceroot.ai 💻 Join our open-source observability & debugging community for more details: GitHub: github.com/traceroot-ai/t… Discord: discord.com/invite/tPyffEZ… #AI #OpenSource #Debugging #Benchmarking #DeveloperTools #GPT #OpenAI #gptoss #ycombinator #gpt5

TraceRoot.AI (YC S25) on the front page of Hacker News 🤯 👉🏼 bit.ly/4mrYk00 Drop by and tell us how you’re debugging production issues today - we'd love to hear from you!

RT @xinwei_97: Appreciate the kind words! 🙏 Been great working with the @Eigent_AI team to keep their multi-agent systems stable in prod -…