[{"data":1,"prerenderedAt":169},["ShallowReactive",2],{"project-49627872":3},{"id":4,"name":5,"line":6,"line_source":7,"title":8,"url":9,"domain":9,"maker":10,"posted_at":11,"year":12,"points":13,"comments":14,"rank_in_year":15,"shown_in_year":16,"top":17,"tier":18,"category":19,"subcategory":20,"format":21,"keywords":22,"reception":28,"sentiment":9,"stance_mix":29,"scored_comments":33,"aspects":34,"strip":36,"post_text":38,"what":39,"said":61,"maker_in_thread":101,"shape":112,"grid":117,"attempts":131,"same_week":132,"maker_shows":160,"maker_hidden":161,"reader_version":162,"card":163},"49627872","Frigade Assist API","Give your AI agent on-screen guides that show users where to click","title","Show HN: Give your AI agent on-screen guides that show users where to click",null,"pancomplex",1788966545,2026,19,8,2068,4366,48,"shown","ai","in-app ai agent tooling","api",[23,24,25,26,27],"ai agent","in-app guide","product onboarding","rag","ai sdk tool","quiet",{"positive":30,"neutral":31,"mixed":32,"critical":32},3,1,0,4,{"idea":35},{"positive":32,"critical":32},[30,32,32,32,37,37,32,37],5,"Hey HN. I'm Christian, one of the founders of Frigade (YC W23). I've noticed that a lot of in-app AI agents struggle to actually understand the products they exist in.\n\nFor instance, let's say a user asks an agent how to do something in a given SaaS product. In an ideal case, maybe that agent replies saying it has a tool to do the task and just automates that work entirely for the user. That's a great outcome.\n\nBut often that's not the case. Maybe there is no tool call for that exact task, or maybe the user's question is best solved by a specific UI workflow or interface. In these cases, many ",{"problem":40,"audience":46,"pricing":9,"motivation":51,"time_to_build":9,"stack":55,"openness":60,"stage":60,"team_size":60},{"value":41,"source":42,"evidence":43},"In-app AI agents struggle to understand the products they live in, and help center docs go stale","maker",[44],{"id":45},"post",{"value":47,"source":48,"evidence":49},"Teams building AI agents inside SaaS products","inferred",[50],{"id":45},{"value":52,"source":42,"evidence":53},"Agents fall back on slow RAG over outdated help docs when they can't automate a task",[54],{"id":45},[56],{"value":57,"source":42,"evidence":58},"Vercel AI SDK",[59],{"id":45},"unknown",{"praise":62,"objections":79,"requests":87,"questions":95,"alternatives":100},[63],{"value":64,"source":65,"evidence":66},"Short positive reactions to the project","commenter",[67,71,75],{"id":68,"by":69,"text":70},"49635171","mahdiharzallah2","Looks interesting",{"id":72,"by":73,"text":74},"49631514","rrrx3","This is nice.",{"id":76,"by":77,"text":78},"49632336","krttherealest","so great!!!",[80],{"value":81,"source":65,"evidence":82},"Commenter says the underlying software should be built cleaner instead of adding layers",[83],{"id":84,"by":85,"text":86},"49630437","doot","What if we just built less shitty software instead of layering more shit on top of shit on top of…",[88],{"value":89,"source":65,"evidence":90},"Use the UI map to improve the product itself",[91],{"id":92,"by":93,"text":94},"49630074","mox1","Could this be utilized to improve the actual UI itself? As a somewhat naive LLM user, I have often found that when coding up projects, when you get into the area of the UI (in…",[96],{"value":97,"source":65,"evidence":98},"Could the UI map be used to improve the actual UI?",[99],{"id":92,"by":93,"text":94},[],{"comments":30,"answered":31,"key":102},[103,106,109],{"id":104,"text":105},"49630484","Yes for sure. We actually use the map we build of the UI + the questions people ask to automatically identify product gap patterns. These are the presented in a dashboard where…",{"id":107,"text":108},"49630689","Very much agree with the sentiment here - I actually even wrote a blog post about this specific issue lately. There's an entirely separate issue of AI slop ruining well-defined…",{"id":110,"text":111},"49631704","Thank you!",{"commenters":113,"max_depth":114,"maker_replies":30,"first_comment_mins":115,"lifespan_hours":116},6,2,139,7,[118,120,122,124,125,127,128,130],[92,32,30,119,93,32],"project\u002Fneutral\u002Fidea",[104,32,32,121,10,31],"project\u002Fpositive\u002Fidea",[84,31,32,123,85,32],"other\u002Fneutral\u002Fnone",[107,31,32,123,10,31],[68,114,37,126,69,32],"project\u002Fpositive\u002Fnone",[72,30,37,126,73,32],[110,30,32,129,10,31],"maker\u002Fpositive\u002Fnone",[76,33,37,126,77,32],[],[133,138,143,147,152,156],{"id":134,"name":135,"points":136,"category":137},"49567437","OpenTrailPaper",424,"hardware",{"id":139,"name":140,"points":141,"category":142},"49610059","Copperhead",250,"dev_tools",{"id":144,"name":145,"points":146,"category":142},"49562219","TERMy",225,{"id":148,"name":149,"points":150,"category":151},"49603546","Stuxnet",179,"security_privacy",{"id":153,"name":154,"points":155,"category":19},"49613068","LLM Attention Visualization",177,{"id":157,"name":158,"points":115,"category":159},"49623398","Fast Polynomials","learning_science",10,false,"v2",{"id":164,"name":5,"line":6,"ai":161,"cat":19,"year":12,"points":13,"comments":14,"rec":28,"s":9,"top":17,"kw":165,"maker":10,"domain":166,"tier":18,"attempts":32,"strip":167,"url":-1,"mix":168,"praise":64,"objection":81,"sub":20},49627872,[23,24,25,26,27],"","30005505",[30,31,32,32],1794036709693]