Subscribe
Sign in
Langfuse
p/langfuseOpen Source LLM Engineering Platform
Start new thread
Clemens Rawert

Langfuse 2.0 - the open source LLM engineering platform

by•

Langfuse is the open source LLM Engineering Platform. It provides observability, tracing, evaluations, prompt management, playground and metrics to debug and improve LLM apps. Langfuse is open. It works with any model, framework and you can export all data.

Add a comment

Replies

Best
Clemens Rawert
Maker
📌
Hi Product Hunt 👋👋👋 I’m Clemens, co-founder of Langfuse (with @marc_klingen and @max_deichmann). We are so excited to be live on Product Hunt again today launching Langfuse 2.0! We originally launched Langfuse right here in late August 2023 and made it to #1 Product of the Day. Earlier this year we were lucky enough to be awarded a Golden Kitty in the AI Infra category for 2023! Product Hunt holds a special place in our heart and company history. Langfuse has come a long way. Last year, we set out to build open source observability and analytics for LLM apps. We saw a need and gap for a dev-focussed, open source and self-hostable product that allows for complex tracing and nesting of LLM and other workloads. We are excited to introduce Langfuse 2.0 today. We have moved from an observability tool to a platform for LLM engineering. The scope is now wider and the core product has matured so much since last year. A few things you should know about: 👣 Tracing: TS & Python (new: decorator) + OpenAI, Llama Index, Langchain, Litellm + more 🤨 Eval Service: Automatically run evals against all new incoming traces 💾 Datasets: Collect, test and run fine-tuning, testing, and golden datasets in Langfuse 💯 Metrics: Dashboards and analytics on cost, latency and quality 🤖 Prompt Management: Version and deploy prompts from within Langfuse 🕹️ LLM Playground: Engineer your prompts with context in Langfuse 🏎️ Export & Fine Tune: Open GET API and csv/JSON exports to build downstream use cases 🚄 Scale: We’ve invested significantly in scaling and resilience as we’ve scaled to thousands of users and handle millions of events a day 🧑‍🤝‍🧑 Community: Join our thousands of users on GitHub Discussions and Discord See for yourself: ⭐ GitHub: https://github.com/langfuse/lang... ⏯️ Demo (includes data, no credit card): https://langfuse.com/demo 📚 Docs: https://langfuse.com/docs We’re excited to launch the all-new Langfuse on Product Hunt today. We will be in the comments the whole day and can’t wait to hear your thoughts & feedback!
Clemens Rawert
Maker
@julietroseb thank you, Juliet :)
Sathish Nagarajan (SNR)
@marc_klingen @max_deichmann @clemo_sf opensourcing LLM engineering platform is great thing in the AI space. Keep up the good work. Congratulations
Oliver Montes
@marc_klingen @max_deichmann @clemo_sf Really amazing product, best in class, Congratulations langfuse team!
Sajida
Congrats on Langfuse 2.0! From observability to a full LLM engineering platform, impressive growth! The features are powerful, catering to user needs. Like snazzystep shoes, Langfuse empowers users to excel. Best wishes on your Product Hunt launch!https://snazzystep.com/wp-admin/
Marc Mengler
💎 Pixel perfection
We were notified automatically about the new launch. It triggered a Slack discussion of our AI infra experts: 'a very nice alternative to langsmith, and it's also cheaper since it isn't per user :+1: It's also open source, so if we don't want to pay for the service we can host it ourselves.' 'We want to remove langchain anyway next week so it kinda fits' Next steps: we check out if the API works for us. Congrats to the Product(hunt)-led growth! :)
Clemens Rawert
Maker
@marc_mengler Thank you, Marc! have a look at the API at https://api.reference.langfuse.com/ + feel free to have your team reach out to me, we're always happy to add endpoints.
Marc Klingen
Maker
@marc_mengler thanks Marc (hi from Marc to Marc) 🙏 Please reach out any time if you have questions or feedback, I'd love to chat! For more context on Langfuse 2.0, the recording of our launch week town hall might be interesting:
Some more links: We have noted down why we build Langfuse here (open, no framework/llm lock-in, great API to build custom workflows around it): https://langfuse.com/why Public roadmap: https://langfuse.com/roadmap Public roadmap discussions: https://langufse.com/ideas Thanks again, appreciate your support!
Fengjiao Peng
Congrats Langfuse team on the milestone!! Been following your progress. Shipping nonstop :)
Marc Klingen
Maker
Thank you @fengjiao_peng1 🚢🚢🚢 We're excited about this release and the evaluations launch today. Really looking forward to the feedback today!
Albert
Congratulations to Clemens, Marc, and Max on the launch of Langfuse 2.0 on Product Hunt! 🎉 It's incredible to see how Langfuse has evolved into a comprehensive LLM Engineering Platform, providing vital tools like tracing, evaluations, metrics, prompt management, and more for developers. The dedication to building an open source platform that caters to the needs of LLM apps is truly commendable. The enhancements in Langfuse 2.0, from new tracing capabilities to prompt management and dataset functionalities, showcase a significant leap in empowering developers to debug and enhance their applications efficiently. I'm curious about the community engagement aspect you mentioned. How has the interaction been with users on GitHub Discussions and Discord influenced the development of Langfuse? Looking forward to exploring Langfuse 2.0 and witnessing the impact it creates in the developer community! 👩‍💻🚀
Marc Klingen
Maker
@mashy thank you! GitHub Discussions have been essential (next to bug reports via GitHub issues) to build a useful product. Our community has been super awesome and you can have a look at all the discussion that has been going on around langfuse here: https://langufse.com/ideas
Viet Le
So great to see you become more holistic! 🤘🏽
Clemens Rawert
Maker
@viet_le 🙏🙏🙏
Marc Klingen
Maker
@viet_le thanks for your support since day -1 🙏
Timon Christiansen
Congrats on the launch! Really enjoying the product since day 1 🚀
Clemens Rawert
Maker
@timon_christiansen Thank you timon and the whole team fastgen!
Marc Klingen
Maker
@timon_christiansen love that you guys actually built your own golang wrapper
Justus Mattern
Congrats guys! Insane shipping speed 🚢🚢🚢
Clemens Rawert
Maker
@justus_mattern let's go!
Marc Klingen
Maker
@justus_mattern thank you! really excited to finally launch many of the new features 🪢
Pablo Jimenez De Parga
Building an LLM solution without langfuse is unreasonable. The product gives visibility on all our traces. The team is always ready to lend a hand as well. Keep building!
Clemens Rawert
Maker
@pablo_jimenez_de_parga Thank you, Pablo. It's been an immense pleasure working for you for such a long time now!
Marc Klingen
Maker
@pablo_jimenez_de_parga your feedback has been invaluable, thanks team throxy!
Mathias Nestler
Probably one of the best LLM platforms out there!!
Clemens Rawert
Maker
@mathias_n Nice to (sort of) run into you yesterday!
Tobias Kemkes
Congrats on the launch guys!! Keep shipping 🚢
Marc Klingen
Maker
@tobiaskemkes thank you, you guys are an inspiration!
Clemens Rawert
Maker
Giveaway: Sign up for Langfuse stickers here: https://langfuse.com/stickers we'll send these in the coming weeks!
Marc Klingen
Maker
@clemo_sf I want some 🪢🪢🪢
Sebastian Schüller
@clemo_sf @marc_klingen @max_deichmann congrats on the launch! It's awesome to follow your journey which is truly lightspeedy. 🔥
Marc Klingen
Maker
@clemo_sf @max_deichmann @sebastian_schuller thanks for your excitement and support, really appreciate it ❤️
Roman Martirosyan
The prompt management and tracing features are particularly intriguing. What measures have you taken to ensure data security and privacy, especially since this is a platform that interacts with potentially sensitive information?
Marc Klingen
Maker
@r_martirosyan data security and privacy is extremely important to us as many traces in Langfuse contain PII or valuable IP of our customers. Short summary: - US / data residency options - Professional penetration test to ensure our interfaces are safe - Following security best practices - GDPR, ISO 27001, and soon SOC 2 Type 2 compliance (pending) Find out more information here https://langfuse.com/security and feel free to reach out to me or security@langfuse.com if you have any questions.
Felix Vemmer
Congrats 👏 on the launch! Excited to try some of these new features!
Marc Klingen
Maker
@felix_vemmer thank you! list of all the latest releases: https://langfuse.com/changelog let me know if you have any feedback!
Aqib Ansari
Really happy with the product. It's very intuitive - Traces work very well for Agentic Workflows - Scores is another great feature to help us create datasets for finetuning - Plus, looking forward to your new eval feature!
Marc Klingen
Maker
@aqib_ansari thanks Aqib, appreciate your continuous product feedback a lot! It really helps us build the best product and continuously learn what your team needs to build the best agents 🙏
Jean Kaddour
This team ships faster than ⚡
Clemens Rawert
Maker
@jean_kaddour thank you, Jean! Likewise with sevn!
Marc Klingen
Maker
@jean_kaddour thanks, Jean, means a lot!
Dan
🔌 Plugged in
We love to use Langfuse to build our RAG systems and get the visibility into what prompts actually hitting the LLM! Btw, the new decorate is awesome
Marc Klingen
Maker
@dan_meier1 thanks, the decorator is for sure the biggest upgrade to the Langfuse developer experience in Python! We wrote a more technical blog post on it here: https://langfuse.com/blog/2024-0...
Niklas Hölterhoff
Love the new features! Great job on shipping so fast!
Marc Klingen
Maker
thanks @nhoelterhofflangdock, you are an inspiration! shipping is all we do, follow along here: https://langfuse.com/changelog
Anastasia Chakileva
Congrats on the launch! I'm curious about how Langfuse handles scaling, especially with the growing volume of events and users. Can you elaborate on the scaling strategies you've implemented to ensure smooth performance as you scale?
Marc Klingen
Maker
@anastasia_chakileva_ we currently scale out our application, on Google Cloud Run and Vercel. We closely measure capacity and performance and this setup scales really well. We are currently working on v3.0 of Langfuse which will include some infrastructure changes to make Langfuse even more scalable, support larger traces with videos/images, and faster dashboard performance.
Robert Lukoshko
best LLM engineering platform I have seen so far. Great job!
Marc Klingen
Maker
@karmedge thanks for your kind words, let us know if you have any feedback! love what you built with Fixkey: https://www.producthunt.com/prod...