Launching today
A workflow automation platform that routes tasks to the right agent – human or AI. By connecting tools like Slack, Jira and Notion – Trace breaks down existing workflows, spots automation opportunities and embeds AI agents to repetitive tasks.
AssemblyAI — Speech-to-Text API with diarization
Speech-to-Text API with diarization
Promoted
Trace
👋 Hey Product Hunt Fam!
I'm Tim, co-founder & ceo of Trace (YC S25), and I’m thrilled to share our public launch 🚀 we previously built @Clustr a web3 analytics platform that grew to 100,000+ users 🙀
👾 What is Trace?
Trace is a workflow automation platform that routes tasks to the right agent – human or AI. By connecting tools like @Slack, @Jira, and @Notion, @Trace breaks down existing workflows and assigns repetitive tasks to AI agents. We help coordinate tasks between agents and people – without disrupting how your teams work.
🧐 Why use Trace?
Routine work like chasing updates, filling docs, and coordinating teams drains time and focus – with Trace you can:
✍️ Generate workflows from a single prompt and track project lifecycles
🧩 Break down tasks into steps and assign the right agents for the job
🤖 Deploy agents for any task using templates or create your own
🚀 Run workflows automatically with real-time triggers and schedules
🦄 How It Works
Trace integrates with your systems to analyze wikis, docs, chats, activity logs, and people data, building a unified index of your company’s work. This provides rich context for every new workflow, making it easy to route tasks across people and AI agents.
🎁 Join us
Ready to ditch the busywork? Sign up for free and tell us which workflows you’ve tried to automate – we’d love to talk.
Email us 📨 founders@trace.so or book a call ☎️
@timcha_cherkasov The unified company index that analyzes all your wikis, docs, and chats to provide rich context for workflow routing is a game-changer - this is exactly what's needed to make AI agents truly effective instead of working in silos. Best of luck with Trace!
Trace
@timcha_cherkasov @kay_arkain Thanks a lot! That unified context layer is exactly what we believe unlocks the next step - an agentic infrastructure that learns from your company and introduces automation without human intervention. A long way to go but you can get the glimpse of it with the current version!
Trace
@timcha_cherkasov @kay_arkain Really appreciate that! We’ve seen how scattered tools make AI feel "broken", even though we know that the tool is available and it should work. Proofs that bringing everything into one context is what makes it click and finally start working.
Trace
@timcha_cherkasov @kay_arkain @zorinvlad under the user interface that non-technical person can understand, of course!
Scade.pro
@timcha_cherkasov looks great. best of luck today
Trace
@timcha_cherkasov @nastassia_k thanks for your support, it means a lot to us
Trace
@timcha_cherkasov @nastassia_k thank you Nastassia 🙏
Trace
@timcha_cherkasov @nastassia_k thanks! 🎊
1Page
@timcha_cherkasov nice twist
@timcha_cherkasov Hey Tim, looks cool! A kind of workflow gallery would be cool, where you can browse use cases where Trace really shines and brings a ton of value (as in: the functionality is really good, and teams are using it successfully).
One specific workflow for which I haven't found a great solution yet: I create a video, generate a title and description, and then I want to post it across different platforms (YouTube, Tiktok, IG, Facebook, Tumblr, X, Pinterest). YouTube is the primary one, so the title and description are optimized for that, and then all the other platforms require kind of different processing (title rewrites, description needs to be rewritten in a platform-specific language, sometimes hashtags generated or pulled from an already existing set of hashtags triggered by certain keywords in the title, etc etc). It's a task that takes us about 20 minutes per video, but it's super repetitive and mindnumbing, yet all the existing tools we've tried so far that claim to do that don't allow for the customization we need or do it in some generic way that kills engagement compared to when we do it ourselves - but it's definitely very formulaic (most of it is just a set of ChatGPT prompts). If that could be automated into a "one click workflow", would be a huge win. You think that's possible with Trace?
Trace
@timcha_cherkasov @startupsam thanks for the thoughtful feedback. The workflow gallery is a good idea. As we learn more about how people use Trace, we will be able to derive a set of common use-cases that we will show on the main page (or maybe on the template store?) so you can easily replicate them.
Your example is a great use case. There is a way to do it now, but it would require some prompt engineering for each of the workflow tasks. It is definitely possible, but now you made me think that there is an opportunity for us to allow video upload to the system so that you can upload a video and get whatever you want from it (transcribe, extract key words, generate tags etc.) 🤔 I will pass it over to the team, thanks a lot for your feedback!
@timcha_cherkasov Congratulations Tim on your second launch on PH. :)
Trace looks super practical especially because it strikes a balance between AI and human. May I ask, how does Trace decide when a task should be routed to an AI agent vs. a human and can teams customize that logic?
Trace
@timcha_cherkasov @rohanrecommends Great question! In Trace, the system has a set of rules depending on the type of flow, number and complexity of tasks and many other parameters to decide what should stay with a human and what can be handled by AI. But the idea of teams tweaking the parameters sounds interesting. It might be nice to have flexibility.
Trace
@timcha_cherkasov @rohanrecommends @arthur_romanov You also have to think about UX here. Too much control can bring confusion. It would be nice if the system works for a general use case out-of-the-box, and parameterisation is a feature for pro users
Trace
@timcha_cherkasov @rohanrecommends You can choose to do the task yourself or assign a teammate if you believe that the task is not good enough for an AI to do. We put users in control. However, we learn from how you use the tool, and it becomes clear with time that "this user doesn't like the task of this type to be automated"
I’ve tried to stitch together Jira + Notion + Slack with zaps and it always breaks. Let's see if Trace can actually keep things running reliably.
Trace
@igorsorokinua thanks for checking Trace out! The PH launch brought a lot of traffic to our site, and some things broke. We are fixing things as we speak.
FuseBase
Trace
@igorsorokinua hey Igor, thanks for the feedback. We would like to improve the feedback loop between user and AI Agent. It may not be very clear when something goes wrong sometimes, but we are working on improving that!
Trace
@igorsorokinua we experience high load during the launch. Please visit the webapp again when you have a chance
FuseBase
Huge congrats on the launch 👏
We run multiple cross-functional projects, and I constantly struggle with visibility. Does Trace give a central “dashboard” of all active workflows?
Trace
@kate_ramakaieva yes, we provide the dashboard where you can track all active workflows. We are working on making it much more comprehensive, like accessing the AI Agent performance, task state trackers with alarms and debug capabilities, task anomaly trackers monitoring things like "how long it has been stale", etc. A lot of new things are coming very soon!
Trace
@kate_ramakaieva, hey Kate, please let us know if there are any specific features you would like to see in the dashboard. Breakdown of AI agent run stats is the one which was requested the most.
Trace
@kate_ramakaieva thanks Kate, you can check state of workflows from the menu right now. As @arthur_romanov said more comprehensive dashboard is coming soon.