[{"data":1,"prerenderedAt":201},["ShallowReactive",2],{"project-42906756":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":40,"post_text":42,"what":43,"said":60,"maker_in_thread":120,"shape":128,"grid":131,"attempts":158,"same_week":163,"maker_shows":41,"maker_hidden":193,"reader_version":194,"card":195},"42906756","omiai","A full mulimodal LLM by merging multiple models into one","title","Show HN: I built a full mulimodal LLM by merging multiple models into one","https:\u002F\u002Fgithub.com\u002FJigsawStack\u002Fomiai","github.com","yoeven",1738480483,2025,40,14,1434,3837,38,"shown","ai","llm routing framework","library",[24,25,26,27],"llm router","multimodal llm","openai compatible","llm framework","roasted",-0.111,{"positive":31,"neutral":32,"mixed":33,"critical":34},2,1,0,3,6,{"idea":37,"execution":38,"design":39},{"positive":33,"critical":34},{"positive":31,"critical":33},{"positive":33,"critical":33},[32,33,33,33,33,33,33,32,41,33,34,41,32,33],5,null,{"problem":44,"audience":42,"pricing":42,"motivation":42,"time_to_build":42,"stack":51,"openness":59,"stage":59,"team_size":59},{"value":45,"source":46,"evidence":47},"Using different LLMs for different strengths requires separate integrations and model selection","inferred",[48],{"id":49,"by":11,"text":50},"42908415","It's a framework that uses the best part of each LLM, e.g. multimodal support from gemini with tool calling from gpt-4o and reasoning from o3-mini by chaining them dynamically.…",[52],{"value":53,"source":54,"evidence":55},"gguf models via llama.cpp (experimental local models branch)","maker",[56],{"id":57,"by":11,"text":58},"42908368","Thank you!! Local model is something I'm looking into with gguf files & llama.cpp, still pretty experimental, you can check out the branch here:…","unknown",{"praise":61,"objections":73,"requests":105,"questions":111,"alternatives":112},[62],{"value":63,"source":64,"evidence":65},"Clean, well-abstracted developer-facing API","commenter",[66,70],{"id":67,"by":68,"text":69},"42908217","upghost","I'll jump in before the haterade engine wakes up -- great bit of engineering work here! I can't imagine a better level of abstracting away the unnecessary stuff while still…",{"id":71,"by":68,"text":72},"42908478","Nice! To be clear for my usecase I didn't mean calling the local LLM directly, rather simply being able to point to an OpenAI compatible API would be fine. It seems like…",[74,97],{"value":75,"source":64,"evidence":76},"Title misleadingly suggests model merging or building a true multimodal LLM; it is actually a router",[77,81,85,89,93],{"id":78,"by":79,"text":80},"42908360","kouteiheika","I clicked expecting a single full multimodal LLM made by merging multiple existing models into one like the title suggests (which sounds very interesting), and I found... a…",{"id":82,"by":83,"text":84},"42908563","danielbln","The problem with that phrasing is that there is actual model merging, where you merge the weights. So people reading the title might (and apparently do) expect that, less so an…",{"id":86,"by":87,"text":88},"42908531","willwade","You know - the word \"multimodal\" i think is being used badly here. Its Multi-Model - not Multimodal - which certainly suggests a completeley different thing",{"id":90,"by":91,"text":92},"42909006","FilipSivak","You clearly don't understand what multimodal means. Multimodal is for example new gemini where you can input green car and get the very same car, only with red paint. Multimodal…",{"id":94,"by":95,"text":96},"42910039","OutOfHere","You are still not getting it. The use of the word multimodal does nothing good for your software. It is an LLM router. I get it that your software does support some multimodal…",{"value":98,"source":64,"evidence":99},"Misuse of terms 'multimodal' and 'merging' by the maker",[100,104],{"id":101,"by":102,"text":103},"42909308","vikramkr","Makes sense but the problem is that you're using words that already have specific meanings in the space, all related to creating one model with multiple functionalities. Merging…",{"id":94,"by":95,"text":96},[106],{"value":107,"source":64,"evidence":108},"Support for local or in-house LLMs, or OpenAI-compatible endpoints",[109,110],{"id":67,"by":68,"text":69},{"id":71,"by":68,"text":72},[],[113],{"value":114,"source":64,"evidence":115},"Langdock",[116],{"id":117,"by":118,"text":119},"42908476","madduci","And a similar product already exists, Langdock",{"comments":121,"answered":32,"key":122},4,[123,124,127],{"id":49,"text":50},{"id":125,"text":126},"42909334","I 100% agree, this simulates a multimodal input and automatically handles the rest along with model selection by using a variation of techniques. It doesn't do this natively on…",{"id":57,"text":58},{"commenters":129,"max_depth":41,"maker_replies":121,"first_comment_mins":130,"lifespan_hours":129},10,310,[132,134,136,138,139,140,142,144,145,147,149,151,155,156],[78,33,32,133,79,33],"project\u002Fcritical\u002Fidea",[49,33,33,135,11,32],"maker\u002Fneutral\u002Fidea",[82,33,33,137,83,33],"commenter\u002Fneutral\u002Fidea",[101,33,33,137,102,33],[125,33,33,135,11,32],[94,33,33,141,95,33],"commenter\u002Fcritical\u002Fidea",[117,33,33,143,118,33],"other\u002Fneutral\u002Fnone",[86,33,32,133,87,33],[67,32,41,146,68,33],"project\u002Fpositive\u002Fexecution",[57,32,33,148,11,32],"project\u002Fneutral\u002Fnone",[71,32,34,150,68,33],"project\u002Fneutral\u002Fdesign",[152,31,41,153,154,33],"42908980","maker\u002Fpositive\u002Fexecution","Peretus",[90,34,32,133,91,33],[157,34,33,137,11,32],"42909318",[159,162],{"id":160,"t":161,"points":32,"comments":33},"42842788",1737995270,{"id":4,"t":12,"points":14,"comments":15},[164,169,174,179,183,188],{"id":165,"name":166,"points":167,"category":168},"42918902","videotoflip",513,"consumer_social",{"id":170,"name":171,"points":172,"category":173},"42901007","ESP32 RC Cars",335,"hardware",{"id":175,"name":176,"points":177,"category":178},"42908061","Lume",309,"dev_tools",{"id":180,"name":181,"points":182,"category":178},"42889407","Uscope",262,{"id":184,"name":185,"points":186,"category":187},"42877399","Audiocube",253,"creative",{"id":189,"name":190,"points":191,"category":192},"42946918","HN hell",170,"productivity",false,"v2",{"id":196,"name":5,"line":6,"ai":193,"cat":20,"year":13,"points":14,"comments":15,"rec":28,"s":197,"top":18,"kw":198,"maker":11,"domain":10,"tier":19,"attempts":31,"strip":199,"url":9,"mix":200,"praise":63,"objection":75,"sub":21},42906756,-0.11,[24,25,26,27],"10000001503510",[31,32,33,34],1793991112081]