[{"data":1,"prerenderedAt":180},["ShallowReactive",2],{"project-47744683":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":34,"aspects":35,"strip":38,"post_text":40,"what":41,"said":67,"maker_in_thread":106,"shape":118,"grid":121,"attempts":142,"same_week":143,"maker_shows":107,"maker_hidden":173,"reader_version":174,"card":175},"47744683","Rekal","Long-term memory for LLMs in a single SQLite file","title","Show HN: Rekal – Long-term memory for LLMs in a single SQLite file","https:\u002F\u002Fgithub.com\u002Fjanbjorge\u002Frekal","github.com","jeeybee",1776029153,2026,9,10,3371,4366,78,"shown","ai","llm memory","cli",[21,24,25,26,27],"mcp server","sqlite","hybrid search","vector search","mixed",0,{"positive":31,"neutral":32,"mixed":29,"critical":33},3,2,1,6,{"execution":36,"design":37},{"positive":33,"critical":33},{"positive":32,"critical":29},[31,39,29,39,29,39,31,29,33,29],5,"I got tired of repeating myself to my LLM every session. rekal is an MCP server that stores memories in SQLite and retrieves them with hybrid search (BM25 + vectors + recency decay). One file, local embeddings, no API keys.",{"problem":42,"audience":48,"pricing":48,"motivation":49,"time_to_build":48,"stack":53,"openness":66,"stage":66,"team_size":66},{"value":43,"source":44,"evidence":45},"LLMs forget context between sessions, so users repeat themselves","maker",[46],{"id":47},"post",null,{"value":50,"source":44,"evidence":51},"Got tired of repeating himself to the LLM every session",[52],{"id":47},[54,62],{"value":55,"source":56,"evidence":57},"Python","commenter",[58],{"id":59,"by":60,"text":61},"47745626","axeldunkel","Nice clean project - Python 3.14 as a requirement might be an issue for adoption...",{"value":63,"source":44,"evidence":64},"SQLite",[65],{"id":47},"unknown",{"praise":68,"objections":80,"requests":85,"questions":97,"alternatives":105},[69,76],{"value":70,"source":56,"evidence":71},"Hybrid search scoring is well thought out",[72],{"id":73,"by":74,"text":75},"47756377","eitanlebras","The hybrid search scoring is really well thought out. One question: are the 0.4\u002F0.4\u002F0.2 weights fixed or configurable? To me it feels like different use cases would want…",{"value":77,"source":56,"evidence":78},"Clean project",[79],{"id":59,"by":60,"text":61},[81],{"value":82,"source":56,"evidence":83},"Python 3.14 requirement may hinder adoption",[84],{"id":59,"by":60,"text":61},[86],{"value":87,"source":56,"evidence":88},"Make hybrid search weights (0.4\u002F0.4\u002F0.2) configurable",[89,90,94],{"id":73,"by":74,"text":75},{"id":91,"by":92,"text":93},"47756660","SachitRafa","All memories should not decay at the same rate. these weights should be configurable !",{"id":95,"by":74,"text":96},"47758266","I think that'd be cool, maybe just a simple config option in pyproject.toml or a .rekal\u002Fconfig.yml would cover most use cases: they could default to 0.4\u002F0.4\u002F0.2 but could be…",[98],{"value":99,"source":56,"evidence":100},"How effective has it been for the maker's own use, and where does it fall over?",[101],{"id":102,"by":103,"text":104},"47757982","alsetmusic","I’m going to try this out. Curious how effective it’s been for your own use, OP? Anything particularly interesting or where it still falls over and requires tuning that you’ve…",[],{"comments":107,"answered":33,"key":108},4,[109,112,115],{"id":110,"text":111},"47757862","The 0.4\u002F0.4\u002F0.2 weights are just based of hand wavy trail an error, will look into other options. I could make them configurable?",{"id":113,"text":114},"47761872","Added a mcp command for it, but i might add it as a comfig as well? https:\u002F\u002Fgithub.com\u002Fjanbjorge\u002Frekal\u002Fpull\u002F14\u002Fchanges",{"id":116,"text":117},"47757864","Thanks, i will look into making them configurable.",{"commenters":39,"max_depth":39,"maker_replies":107,"first_comment_mins":119,"lifespan_hours":120},128,54,[122,124,126,128,130,132,134,136,138,140],[102,29,31,123,103,29],"maker\u002Fneutral\u002Fnone",[73,33,39,125,74,29],"project\u002Fpositive\u002Fexecution",[110,33,29,127,11,33],"maker\u002Fneutral\u002Fexecution",[95,33,39,129,74,29],"project\u002Fpositive\u002Fdesign",[113,33,29,131,11,33],"maker\u002Fpositive\u002Fdesign",[133,33,39,131,74,29],"47774241",[91,32,31,135,92,29],"project\u002Fneutral\u002Fdesign",[116,32,29,137,11,33],"maker\u002Fneutral\u002Fdesign",[59,31,33,139,60,29],"project\u002Fcritical\u002Fexecution",[141,31,29,127,11,33],"47748255",[],[144,149,154,159,164,168],{"id":145,"name":146,"points":147,"category":148},"47742200","boringBar",520,"productivity",{"id":150,"name":151,"points":152,"category":153},"47727960","Pardonned.com",501,"data",{"id":155,"name":156,"points":157,"category":158},"47739174","Oberon System 3",245,"infrastructure",{"id":160,"name":161,"points":162,"category":163},"47778306","tracksuccession.com",191,"web_saas",{"id":165,"name":166,"points":167,"category":163},"47749674","BrightBean Studio",190,{"id":169,"name":170,"points":171,"category":172},"47705413","Craft",177,"dev_tools",false,"v2",{"id":176,"name":5,"line":6,"ai":173,"cat":20,"year":13,"points":14,"comments":15,"rec":28,"s":29,"top":18,"kw":177,"maker":11,"domain":10,"tier":19,"attempts":29,"strip":178,"url":9,"mix":179,"praise":70,"objection":82,"sub":21},47744683,[21,24,25,26,27],"3505053010",[31,32,29,33],1793991151267]