[{"data":1,"prerenderedAt":334},["ShallowReactive",2],{"project-37310070":3},{"id":4,"name":5,"line":6,"line_source":7,"title":8,"url":9,"domain":10,"maker":11,"posted_at":12,"year":13,"points":14,"comments":15,"rank_in_year":16,"shown_in_year":17,"top":18,"tier":19,"category":20,"subcategory":21,"format":22,"keywords":23,"reception":28,"sentiment":29,"stance_mix":30,"scored_comments":35,"aspects":36,"strip":42,"post_text":44,"what":45,"said":104,"maker_in_thread":201,"shape":213,"grid":217,"attempts":298,"same_week":299,"maker_shows":33,"maker_hidden":326,"reader_version":327,"card":328},"37310070","Langfuse","Open-source observability and analytics for LLM apps","title","Show HN: Langfuse – Open-source observability and analytics for LLM apps","https:\u002F\u002Fgithub.com\u002Flangfuse\u002Flangfuse","github.com","marcklingen",1693325646,2023,143,35,443,3109,15,"notable","dev_tools","llm observability","web_app",[21,24,25,26,27],"llm tracing","llm analytics","open source","llm monitoring","curious",0.217,{"positive":31,"neutral":32,"mixed":33,"critical":34},6,14,1,2,23,{"idea":37,"execution":39,"design":41},{"positive":38,"critical":34},3,{"positive":33,"critical":40},0,{"positive":40,"critical":40},[43,38,38,38,40,38,43,40,38,38,38,40,38,38,43,38,40,38,40,40,40,40,38,43,40,33,40,38,38,43,38,33,43],5,"Hi HN! Langfuse is OSS observability and analytics for LLM applications (repo: https:\u002F\u002Fgithub.com\u002Flangfuse\u002Flangfuse, 2 min demo: https:\u002F\u002Flangfuse.com\u002Fvideo, try it yourself: https:\u002F\u002Flangfuse.com\u002Fdemo)\n\nLangfuse makes capturing and viewing LLM calls (execution traces) a breeze. On top of this data, you can analyze the quality, cost and latency of LLM apps.\n\nWhen GPT-4 dropped, we started building LLM apps – a lot of them! [1, 2] But they all suffered from the same issue: it’s hard to assure quality in 100% of cases and even to have a clear view of user behavior. Initially, we logged all prompts",{"problem":46,"audience":52,"pricing":60,"motivation":64,"time_to_build":68,"stack":69,"openness":101,"stage":102,"team_size":103},{"value":47,"source":48,"evidence":49},"Hard to assure quality and understand user behavior in LLM apps","maker",[50],{"id":51},"post",{"value":53,"source":48,"evidence":54},"Teams building LLM applications",[55,56],{"id":51},{"id":57,"by":58,"text":59},"37314739","mdeichmann","Thanks a lot! We see teams adopt Langfuse quite early already. Say you have one or two engineers working on a rather complex LLM feature, they look for a solution like Langfuse…",{"value":61,"source":48,"evidence":62},"Paid managed cloud version; self-hosting free with enterprise features reserved",[63],{"id":51},{"value":65,"source":48,"evidence":66},"Needed better context and analytics to improve LLM apps after building them",[67],{"id":51},null,[70,74,78,82,86,90,94],{"value":71,"source":48,"evidence":72},"TypeScript",[73],{"id":51},{"value":75,"source":48,"evidence":76},"Next.js",[77],{"id":51},{"value":79,"source":48,"evidence":80},"Prisma",[81],{"id":51},{"value":83,"source":48,"evidence":84},"tRPC",[85],{"id":51},{"value":87,"source":48,"evidence":88},"Tailwind",[89],{"id":51},{"value":91,"source":48,"evidence":92},"NextAuth",[93],{"id":51},{"value":95,"source":48,"evidence":96},"Postgres on Supabase",[97,98],{"id":51},{"id":99,"by":11,"text":100},"37312688","Currently the project is all Postgres and the managed version is on Supabase. No scalability issues yet, we consider different OLAP options as we are mostly interested in faster…","open_source","launched","team",{"praise":105,"objections":130,"requests":147,"questions":164,"alternatives":180},[106,122],{"value":107,"source":108,"evidence":109},"Open source and clear value proposition","commenter",[110,114,118],{"id":111,"by":112,"text":113},"37310465","steventey","> We will need to move to an OLAP database soon and are debating if we need to start batching ingestion Highly recommend https:\u002F\u002Ftinybird.com for this – they're a fantastic OLAP…",{"id":115,"by":116,"text":117},"37313753","addisonj","Congrats on the launch! I have quite a few years of observability experience behind me and hand't really considered some of the unique aspects that LLMs bring into the picture.…",{"id":119,"by":120,"text":121},"37312245","ij23","looks awesome ! love that it's Open-source too",{"value":123,"source":108,"evidence":124},"LLM-specific observability is a real underserved need",[125,129],{"id":126,"by":127,"text":128},"37313608","elamje","Awesome. There is a definitely a need for LLM product analytics that is currently completely underserved by traditional tools like GA, Mixpanel, etc.",{"id":115,"by":116,"text":117},[131,143],{"value":132,"source":108,"evidence":133},"Should support OpenTelemetry rather than a custom SDK",[134,138,142],{"id":135,"by":136,"text":137},"37313140","phillipcarter","Congrats on the release! I'm keenly interested in this space, as I believe that Observability is one of the top ways to steer LLMs to be more reliable in production. I noticed…",{"id":139,"by":140,"text":141},"37314866","hrpnk","+1 on the OTel mention. Having telemetry in place in a system, one would typically implement a single behavioral tracking SDK on top. Adding yet another SDK for LLMs is a hard…",{"id":115,"by":116,"text":117},{"value":144,"source":108,"evidence":145},"Pricing by storage may cause cost variance versus events per second",[146],{"id":115,"by":116,"text":117},[148,156],{"value":149,"source":108,"evidence":150},"OpenTelemetry support",[151,152],{"id":135,"by":136,"text":137},{"id":153,"by":154,"text":155},"37315050","anirudhrx","Congrats on the launch! This is really cool. Would love to see OTel integration in the future. I'm curious if this might eventually work with request-context based routing, i.e.…",{"value":157,"source":48,"evidence":158},"Replace Looker Studio with an open-source analytics option",[159,160],{"id":51},{"id":161,"by":162,"text":163},"37313168","fiehtle","If you’re looking to replace Looker with open source and the ability to style it to your needs maybe a mix of cube.dev plus tremor.so would do the trick?",[165,172],{"value":166,"source":108,"evidence":167},"How does it compare to Datadog LLM monitoring?",[168],{"id":169,"by":170,"text":171},"37311659","pranay01","Congrats on the launch! Curious to learn what specific use case you have seen around observability of LLM apps which are not covered by standard observability tools like DataDog,…",{"value":173,"source":48,"evidence":174},"Timescale or ClickHouse as the OLAP choice?",[175,176],{"id":51},{"id":177,"by":178,"text":179},"37312121","kaspermarstal","I’m curious if you investigated the TimescaleDB extension that is built into Supabase for your usecase? And if so, what was the pros and cons?",[181,185,189,193,197],{"value":182,"source":108,"evidence":183},"Datadog",[184],{"id":169,"by":170,"text":171},{"value":186,"source":108,"evidence":187},"SigNoz",[188],{"id":169,"by":170,"text":171},{"value":190,"source":48,"evidence":191},"Lightdash",[192],{"id":51},{"value":194,"source":108,"evidence":195},"Cube.dev with tremor",[196],{"id":161,"by":162,"text":163},{"value":198,"source":108,"evidence":199},"Tinybird",[200],{"id":111,"by":112,"text":113},{"comments":202,"answered":33,"key":203},4,[204,207,210],{"id":205,"text":206},"37315378","Thx. Currently looking deeply into how we can make it as dynamic as possible to help people experiment with configurations and test on production samples. What do you exactly…",{"id":208,"text":209},"37321604","Many great points\u002Fideas here and on Discord, thanks HN! For those reading this thread later, feel free to reach out with any feedback or questions marc at langfuse dot com",{"id":211,"text":212},"37312328","Thanks for the hint. Having only one fully managed db seems to be interesting to reduce needs for joining across databases and less operations overhead (managing uptime, setting…",{"commenters":214,"max_depth":31,"maker_replies":202,"first_comment_mins":215,"lifespan_hours":216},18,28,33,[218,220,223,226,229,233,237,238,239,241,244,245,247,249,252,253,256,257,259,260,264,267,269,272,275,277,279,283,285,287,290,292,295],[135,40,43,219,136,40],"project\u002Fpositive\u002Fidea",[221,40,38,222,58,40],"37313654","maker\u002Fneutral\u002Fexecution",[224,40,38,225,136,40],"37314197","project\u002Fneutral\u002Fidea",[227,40,38,228,58,40],"37314307","maker\u002Fneutral\u002Fidea",[230,40,40,231,232,40],"37317087","other\u002Fneutral\u002Fnone","eoproc",[234,33,38,235,236,40],"37316797","project\u002Fneutral\u002Fnone","jayunit",[153,34,43,219,154,40],[205,34,40,228,11,33],[240,34,38,228,154,40],"37315495",[242,38,38,225,243,40],"37314590","v3np",[57,38,38,225,58,40],[208,202,40,246,11,33],"maker\u002Fneutral\u002Fnone",[161,43,38,248,162,40],"project\u002Fneutral\u002Fdesign",[250,43,38,251,58,40],"37313741","maker\u002Fneutral\u002Fdesign",[126,31,43,219,127,40],[177,254,38,255,178,40],7,"project\u002Fneutral\u002Fexecution",[211,254,40,222,11,33],[258,254,38,222,178,40],"37312520",[99,254,40,222,11,33],[261,254,40,262,263,40],"37313511","other\u002Fneutral\u002Fexecution","zacmps",[265,254,40,266,178,40],"37318825","commenter\u002Fneutral\u002Fexecution",[268,254,40,266,263,40],"37331253",[115,270,38,271,116,40],8,"project\u002Fmixed\u002Fidea",[273,270,43,274,58,40],"37314638","maker\u002Fpositive\u002Fnone",[276,270,40,262,116,40],"37316145",[139,270,33,278,140,40],"project\u002Fcritical\u002Fidea",[280,270,40,281,282,40],"37317062","commenter\u002Fpositive\u002Fnone","ecnahc515",[169,284,38,225,170,40],9,[286,284,38,228,58,40],"37312299",[111,288,43,289,112,40],10,"project\u002Fpositive\u002Fexecution",[291,288,38,222,58,40],"37311408",[293,288,33,278,294,40],"37312084","NathanFlurry",[119,296,43,297,120,40],11,"project\u002Fpositive\u002Fnone",[],[300,304,309,314,318,321],{"id":301,"name":302,"points":303,"category":20},"37326870","compress",815,{"id":305,"name":306,"points":307,"category":308},"37247767","obsidian-sync",416,"infrastructure",{"id":310,"name":311,"points":312,"category":313},"37252231","Shimmer",229,"consumer_social",{"id":315,"name":316,"points":317,"category":20},"37251969","Continue",187,{"id":319,"name":320,"points":14,"category":308},"37263473","Release AI",{"id":322,"name":323,"points":324,"category":325},"37274882","Crumb",138,"languages",false,"v2",{"id":329,"name":5,"line":6,"ai":326,"cat":20,"year":13,"points":14,"comments":15,"rec":28,"s":330,"top":18,"kw":331,"maker":11,"domain":10,"tier":19,"attempts":40,"strip":332,"url":9,"mix":333,"praise":107,"objection":132,"sub":21},37310070,0.22,[21,24,25,26,27],"533303503330335303000035010335315",[31,32,33,34],1794036640908]