{"block_close_oss.c33fd1fe9c":"Community and ecosystem","core-extras.96d330537a":"Base IR","core-extras.bee831a169":"Substrate","core-extras.61a0822826":"Apps on Core","core-extras.15438cc098":"not supported","core-extras.ad565d9d01":"Feature","core-extras.0155bf03e5":"Segment legend","core-extras.8b7110078e":"Algorithm groups by data type","core-extras.25ab4276a6":"Scrollable capability matrix","open-source-hedl.4bd21f0136":"Multi-agent orchestration with HEDL throughout.","open-source-hedl.485a04fef4":"Nexus","open-source-hedl.fe0a091fdb":"Products","open-source-hedl.3cbe0c2efd":"Sub-millisecond retrieval for richer prompts.","open-source-hedl.c91597c8b9":"BitWeave","open-source-hedl.6cc3364a24":"MEASURED","open-source-hedl.d0d092fb05":"Content-addressed parsing pipeline.","open-source-hedl.188fb5d6cf":"Reed","open-source-hedl.d5d6e83bf2":"Continue exploring","open-source-hedl.75e393a1ac":"Explore related projects","open-source-hedl.f9743247fa":"Read the documentation","open-source-hedl.9393a6cb9d":"Install HEDL","open-source-hedl.dd83fae7a1":"Inspect the benchmark","open-source-hedl.cf033f720c":"Install HEDL, run the MCP server in front of your tools, and watch token usage drop. Benchmark corpus in the repository.","open-source-hedl.8512537945":"Half the tokens. Same data","open-source-hedl.90e40d5043":"Get started","open-source-hedl.d007bb1b85":"HEDL","open-source-hedl.650fabe787":"Data formats","open-source-hedl.46ccad8d31":"CONTENT-ADDRESSED PARSING","open-source-hedl.1db44abe76":"RETRIEVAL","open-source-hedl.e566c66130":"HEDL / COMPACT EXCHANGE","open-source-hedl.818c05f658":"Data stays local","open-source-hedl.13f71ffcde":"No external service","open-source-hedl.b250355010":"Same semantics, denser encoding, restored exactly.","open-source-hedl.f3b36f28ca":"Lossless","open-source-hedl.cd3d2c0f50":"Remove the proxy and you are back on JSON.","open-source-hedl.86b6e48353":"Reversible","open-source-hedl.a97eb63329":"Rust core with zero dependencies and no unsafe code.","open-source-hedl.fb85566784":"Supply chain","open-source-hedl.375baff7ee":"No vendor in path. Conversion runs locally.","open-source-hedl.c10e83c453":"Data residency","open-source-hedl.192e3eadca":"How it sits in your stack","open-source-hedl.273e66d878":"DEPLOYMENT POSTURE","open-source-hedl.9be34046ca":"LOCAL","open-source-hedl.c968638a6e":"LOCAL CONVERSION","open-source-hedl.7e633b78ed":"HEDL ships as a Rust library, MCP server, and WASM component. The core parser has zero dependencies and no unsafe code; conversion is local, bidirectional, and lossless across the supported formats.","open-source-hedl.2f5c74a824":"Same data. Half the cost","open-source-hedl.a55e5a1229":"Measured deployment","open-source-hedl.fca77b7e3a":"The measured case","open-source-hedl.f050358edc":"Lower bill","open-source-hedl.3dac09587b":"Same data","open-source-hedl.c3d66d590b":"Library, MCP server, and WASM expose the same lossless conversion path.","open-source-hedl.fef8e889dc":"Interfaces","open-source-hedl.6b5eb5323e":"Library · MCP · WASM","open-source-hedl.38b1adfcd7":"Drop-in server, zero app change.","open-source-hedl.327a55f82d":"Start with one workflow","open-source-hedl.21593b807a":"MCP","open-source-hedl.4ece7bd466":"On extraction, HEDL versus JSON.","open-source-hedl.12a3a4f498":"Accuracy","open-source-hedl.3a450ee308":"Average across 7 benchmark datasets.","open-source-hedl.abda91dcb0":"Token cut","open-source-hedl.ac2d313c06":"Project stats at a glance","open-source-hedl.8ec9f20dc2":"WHAT SHIPS","open-source-hedl.b6cca12605":"BATCH","open-source-hedl.21e5d14e6d":"EXTRACT","open-source-hedl.1879da5549":"RAG","open-source-hedl.31e78faee8":"AGENTS","open-source-hedl.5f9bc6e2f9":"Agent pipelines, retrieval, batch extraction, anywhere LLM tokens are billed by volume. The savings compound as tool call chains grow longer. Same data goes in, same answers come out, the only thing that changes is the number on the invoice.","open-source-hedl.9511481b3f":"are real money","open-source-hedl.67d1c93608":"Where tokens","open-source-hedl.14a29ba383":"Use cases","open-source-hedl.8951252018":"Lossless round-trip in either direction","open-source-hedl.3e940b17d8":"JSON to HEDL","open-source-hedl.0c3204f14f":"Schema parser, positional encoder, MCP adapter","open-source-hedl.2317d66e2e":"HEDL engine","open-source-hedl.55acb25786":"Browser and edge","open-source-hedl.777bbad8db":"Embed in Rust services","open-source-hedl.6560efcad6":"hedl (crate)","open-source-hedl.7298ed49c0":"Convert, validate, lint, format","open-source-hedl.e5d42d6370":"Front of any tool, zero app change","open-source-hedl.80e2b52ec4":"One engine, four front ends","open-source-hedl.60aacf7cd4":"SHIPPED SURFACES","open-source-hedl.4e946026a7":"WASM","open-source-hedl.803a250cd3":"CRATE","open-source-hedl.700b401ca5":"CLI","open-source-hedl.c1120fc325":"Drop in the MCP server and your agent stack speaks HEDL upstream and JSON downstream with no application code change. The proxy logs token counts before and after on every request. A Rust crate, a CLI, and a WASM build cover embedded and browser use.","open-source-hedl.eb1d7ce983":"Library. WASM","open-source-hedl.c281ad43c8":"MCP server","open-source-hedl.899e5920ef":"JSON outside, HEDL inside","open-source-hedl.7ad55df34d":"Measured on 571 extraction tasks across 7 datasets, run on Mistral Large, DeepSeek Chat, and NVIDIA GLM-4.7.","open-source-hedl.c3d33240c4":"Tasks benchmarked","open-source-hedl.6e3aac9200":"Answers per token","open-source-hedl.ff14e25f15":"Accuracy gain","open-source-hedl.d5bce505c9":"Tokens cut","open-source-hedl.a429d45763":"HEDL at a glance","open-source-hedl.fff438fd89":"You pay for it","open-source-hedl.9830566e4b":"Syntax overhead","open-source-hedl.ca7d769502":"The values you meant to send. In HEDL, nearly all that remains.","open-source-hedl.48cd18dabc":"Your data","open-source-hedl.e231b6fb73":"A quote pair around every string, even the unambiguous ones.","open-source-hedl.7b0257ce75":"Quotes","open-source-hedl.2f4219f749":"Braces and brackets marking structure the schema already knows.","open-source-hedl.19ab36d21e":"Brackets","open-source-hedl.d48a7a265d":"The same field names, paid again on every record.","open-source-hedl.9bc20846df":"Repeated keys","open-source-hedl.774e1f7a77":"The four kinds of bytes in a record","open-source-hedl.ac286719e0":"WHERE TOKENS GO","open-source-hedl.082b45a2cd":"PURE DATA","open-source-hedl.171e8168d9":"QUOTES","open-source-hedl.9387ece353":"BRACKETS","open-source-hedl.fb0d661da7":"REPEATED KEYS","open-source-hedl.232e117357":"JSON is the universal exchange format, but as an LLM input it is wasteful. Repeated field names, structural brackets, and quote marks around every string are bytes you pay for that carry no meaning. Isolate each part below to see exactly what your token budget is buying.","open-source-hedl.4f5a49c257":"syntax, not data","open-source-hedl.e3d3ff4b57":"JSON pays for","open-source-hedl.8a9400005c":"JSON spends tokens on syntax","open-source-hedl.57c9bd6a9c":"Measured outcome","open-source-hedl.06df5bc9d1":"Round-trip JSON to HEDL and back.","open-source-hedl.5d60c7dd62":"Compatibility","open-source-hedl.2c5fb7428e":"HEDL 0.12 versus JSON 0.05.","open-source-hedl.4a83075614":"Answers per 1K tokens","open-source-hedl.0debe90b06":"HEDL 80.4% versus JSON 70.1%.","open-source-hedl.71c850e5f4":"Average across the 7 benchmark datasets.","open-source-hedl.813d9f140d":"Token reduction","open-source-hedl.9bb03f0317":"Average across 3 models and 7 datasets","open-source-hedl.df405937a3":"BENCH RESULTS","open-source-hedl.257edfbafc":"571 TASKS","open-source-hedl.166db0fad0":"DEEPSEEK","open-source-hedl.add65ff97f":"MISTRAL LARGE","open-source-hedl.0dbc4f1aee":"We did not assume LLMs would understand a denser format. We tested it. In one measured run across 571 structured extraction questions on 7 real-world datasets, HEDL used 56 percent fewer tokens than JSON and recorded the highest accuracy among the formats in that run. The result belongs to that model, corpus and test setup.","open-source-hedl.6fd4a41947":"Seven datasets. 571 tasks","open-source-hedl.fb82396800":"Six formats","open-source-hedl.9ed99e75a9":"Benchmark","open-source-hedl.d5a3cabacb":"Fewer tokens","open-source-hedl.2aa16293bd":"JSON to HEDL and back, byte for byte.","open-source-hedl.592789518f":"Round trip","open-source-hedl.21292d8032":"Same records, schema declared once.","open-source-hedl.fa2affc1af":"HEDL tokens","open-source-hedl.f0610d0a94":"Flat user records, per question.","open-source-hedl.5b033236cd":"JSON tokens","open-source-hedl.58f5439c22":"Tokens per question, real datasets","open-source-hedl.9a1243353c":"MEASURED SAVING","open-source-hedl.32e74a6072":"LOSSLESS","open-source-hedl.be8a358f6a":"POSITIONAL","open-source-hedl.eaa4b6eae9":"SCHEMA ONCE","open-source-hedl.7e04bd3452":"HEDL declares a schema once, then encodes records positionally. No repeated field names. No brackets for nesting. Strings unquoted by default. Pick a dataset and watch the token count fall while the data stays identical. The round-trip is lossless.","open-source-hedl.74398797f0":"Half the tokens","open-source-hedl.a4dc1dd2c3":"Same data","open-source-hedl.165e234f49":"The format","open-source-hedl.57f4be9410":"Designed and maintained in Europe, with conversion running wherever you deploy it.","open-source-hedl.cfab3eb33f":"European origin","open-source-hedl.17610e5967":"The Rust core has zero dependencies and no unsafe code.","open-source-hedl.a3b23f1329":"It does not phone home. Nothing crosses a border you did not approve.","open-source-hedl.fc91b1b92f":"No foreign call","open-source-hedl.08143af38c":"No vendor in the path. Conversion runs locally on your own machines.","open-source-hedl.222bc10cb1":"SOVEREIGNTY POSTURE","open-source-hedl.ea9988685b":"EU FOUNDATION","open-source-hedl.71a8857309":"NO EGRESS","open-source-hedl.5b6abcc74c":"Conversion runs on your own machines, in an EU region, or fully offline. HEDL does not phone home and needs no external service, so structured records stay inside the boundary you already operate.","open-source-hedl.3c270abe69":"Stays on your ground","open-source-hedl.4d142242a0":"Runs on your ground","open-source-hedl.9fd8900bba":"Where your data stays","open-source-hedl.0b7ad5baf7":"No big-bang","open-source-hedl.220abe579b":"Remove the proxy and every downstream system continues on plain JSON.","open-source-hedl.33cdfc7cb2":"Keep or leave","open-source-hedl.d78a4e6a05":"The same drop-in, wherever it pays.","open-source-hedl.78b88b8dbd":"Roll it out","open-source-hedl.4a799b063a":"Real token counts, logged per request.","open-source-hedl.d7026fe2d0":"See before and after","open-source-hedl.1f1e8641af":"One AI workflow, nothing else touched.","open-source-hedl.223cab364a":"Try one workflow","open-source-hedl.bc16cd75fb":"No code at any point","open-source-hedl.fc2c6297a4":"THE FOUR STEPS","open-source-hedl.9ffbbcc019":"REVERSIBLE","open-source-hedl.6251fd85d4":"EXPAND","open-source-hedl.0e81bff68c":"MEASURE","open-source-hedl.ba7c70b77b":"PILOT","open-source-hedl.fdb44d735a":"There is no big-bang migration to plan. Put the converter in front of one AI workflow, compare the logged token counts, then expand only where the saving holds. The path is reversible: remove the proxy and every downstream system continues on plain JSON.","open-source-hedl.29eaa46c51":"prove it, then decide","open-source-hedl.21dcb94f7a":"Start small,","open-source-hedl.be473b6765":"How a rollout goes","open-source-hedl.735f111f70":"In and out","open-source-hedl.dee1038d97":"Seven formats","open-source-hedl.fd64f11810":"Remove the converter; databases and reports keep receiving ordinary JSON.","open-source-hedl.4d9684f303":"Clean exit","open-source-hedl.04da594dac":"It converts at the edge, so your databases and reports never change.","open-source-hedl.1e41ef2a5f":"No re-platforming","open-source-hedl.9d9f501b4f":"Parquet for data lakes and Neo4j Cypher for graph databases, both directions.","open-source-hedl.cc1820f293":"The data formats","open-source-hedl.a989c44476":"JSON, YAML, XML and CSV, the shapes most teams already exchange every day.","open-source-hedl.cb1142fc61":"The everyday formats","open-source-hedl.5461d80774":"Both ways, into and out of HEDL","open-source-hedl.3cb1d21212":"WHAT IT SPEAKS","open-source-hedl.b174e1fadd":"PARQUET","open-source-hedl.3281188351":"CSV","open-source-hedl.c0aa9ef764":"XML","open-source-hedl.ffbcb202b0":"YAML","open-source-hedl.031a4e76f0":"JSON","open-source-hedl.8840fbfafe":"Nobody has to rip anything out. HEDL converts to and from the formats your systems already use, in both directions, and sits in front of your AI tools as a single drop-in. The dense version only exists on the trip to the model. Everything downstream sees the shapes it always did.","open-source-hedl.9da852e09d":"what your systems speak","open-source-hedl.a3ebde8f9c":"It speaks","open-source-hedl.d5cf5a4de1":"One format, every system","open-source-hedl.dc99d54d99":"Local","open-source-hedl.27a1f07e6c":"Lower invoice","open-source-hedl.b4f72b77f4":"The meter scales the same rate up and down, so you can read off a figure close to your real usage.","open-source-hedl.2c0d187cdc":"Your own volume","open-source-hedl.9b4a5aed1b":"The saving is simply the tokens HEDL removes. Less to send means less to pay for, at the same data.","open-source-hedl.e19a805469":"56 percent fewer tokens","open-source-hedl.20f2029ee4":"The published reference: a service at one billion tokens a month saves 1,680 dollars a month by switching from JSON.","open-source-hedl.5af8c0c547":"1B tokens a month","open-source-hedl.9655c43156":"Anchored to the published number","open-source-hedl.d8538522db":"HOW THE FIGURE IS BUILT","open-source-hedl.cadd2bc4f1":"NO RE-PLATFORM","open-source-hedl.5a3c580998":"YEARLY","open-source-hedl.d717de9444":"MONTHLY","open-source-hedl.343b9c8948":"56 PERCENT FEWER","open-source-hedl.02437a3619":"Tokens are billed by volume, and HEDL sends 56 percent fewer of them for the same data. Choose the size that looks like your usage and watch the monthly and yearly figures change. The published reference point is a service at one billion tokens a month, which saves 1,680 dollars a month.","open-source-hedl.e3cdb9e89f":"see the bill fall","open-source-hedl.53191aa8c7":"Pick your volume,","open-source-hedl.af3ffca3a2":"Put a number on it","open-source-hedl.3f8ea04804":"Same data back, byte for byte.","open-source-hedl.5cb89150fd":"More usable answers per spend.","open-source-hedl.fb5d8b22ca":"HEDL versus JSON on extraction.","open-source-hedl.2e038a433d":"Higher in this run","open-source-hedl.c874349621":"Measured, not estimated","open-source-hedl.3de6c061e6":"THE NUMBERS","open-source-hedl.79bcc2d134":"+10.3PP ACCURACY","open-source-hedl.9fa6deba51":"56 PERCENT FEWER TOKENS","open-source-hedl.8b602db152":"Every AI call is priced by the token, and much of that budget is syntax rather than data. HEDL sent the same request in 56 percent fewer tokens in the measured run and showed a +10.3pp accuracy difference from JSON there. That is evidence from this setup, alongside a lower bill, without changing downstream systems.","open-source-hedl.0e7e12d88f":"with less to go wrong","open-source-hedl.761d9aba73":"A lower bill,","open-source-hedl.676ffa603b":"What it saves, what it de-risks","open-source-hedl.12df5bb228":"Yours to keep","open-source-hedl.04e627eb01":"Benchmark corpus and round-trip checks make the effect visible before production.","open-source-hedl.7a60981833":"Measured behaviour","open-source-hedl.3b8d62ef4f":"It speaks plain JSON to the rest of your stack. Remove it and you are back on JSON, byte for byte.","open-source-hedl.c0af8906ae":"One compact layer sits before the model; the rest of your systems continue speaking JSON.","open-source-hedl.9aa310f1f8":"Three operating boundaries","open-source-hedl.2507c7f79b":"HOW IT FITS","open-source-hedl.3115880727":"CLEAN EXIT","open-source-hedl.09ecce99f5":"JSON IN AND OUT","open-source-hedl.8fd086392c":"HEDL adds one compact exchange layer before the model while every database, report, and application continues speaking JSON. Run conversion locally, measure it on one workflow, and remove it without a data migration if the numbers do not hold.","open-source-hedl.00d756e2ca":"without changing your stack","open-source-hedl.3d1e7bdec2":"Send less overhead","open-source-hedl.95d71ca3a6":"See how it fits","open-source-hedl.f264604373":"In the background","open-source-hedl.fe73d79fdc":"Quiet","open-source-hedl.126800ea74":"Fewer words to pay for, with the same answer at the end.","open-source-hedl.f706e5551a":"A smaller bill","open-source-hedl.43772ba8b9":"Every compact message is restored and checked before the answer reaches you.","open-source-hedl.7bd0e2f849":"Round trip checked","open-source-hedl.a6ea4a2aca":"It only changes how the message is folded for sending.","open-source-hedl.e307ea77c5":"Your words stay yours","open-source-hedl.20a566aa51":"You never see HEDL or touch it. It sits behind the helper.","open-source-hedl.bbbf622b20":"Nothing to learn","open-source-hedl.8ba90c720a":"Four short lines","open-source-hedl.4e04a9e144":"THE PROMISE","open-source-hedl.b396600e58":"EUROPEAN","open-source-hedl.33906a7dfd":"LOSSLESS","open-source-hedl.b00a146670":"SAME ANSWER","open-source-hedl.bd4e9c0810":"NO SETUP","open-source-hedl.8f187b26fb":"HEDL is one small piece that makes AI helpers cheaper to run, without changing what they say back to you. You do not install it, you do not set it up, and you do not have to understand it. Here is the whole promise, in four short lines.","open-source-hedl.dee00cc586":"thinking about this part","open-source-hedl.aa3d2c9dfd":"You can stop","open-source-hedl.4e51dfc189":"Nothing to worry about","open-source-hedl.fd1cdd3829":"Nothing hidden","open-source-hedl.fffbce7dc4":"Ask anything","open-source-hedl.feecf38ed8":"No. The full message is restored before it is used.","open-source-hedl.da6818fb20":"Nothing goes missing","open-source-hedl.7de1fc6b05":"The full version comes back word for word.","open-source-hedl.4352325231":"Always reversible","open-source-hedl.eadc55ea2a":"The meaning is kept exactly the same.","open-source-hedl.3477f4d91d":"Same answer","open-source-hedl.f0078374f1":"You never install it or switch it on.","open-source-hedl.0b049b5346":"Nothing to do","open-source-hedl.d560f943cd":"If you only read these","open-source-hedl.f42015315c":"THE SHORT ANSWERS","open-source-hedl.a3115ab87f":"It is normal to have questions, and there are no silly ones. Here are the five people ask most often, each with a plain answer and nothing hidden behind it. Tap a question to read the answer.","open-source-hedl.74dff2db65":"people ask first","open-source-hedl.4a367295b6":"The things","open-source-hedl.d0f7666b6b":"Your questions, answered","open-source-hedl.7fa8db1cd1":"It does the folding","open-source-hedl.73867ad514":"You do the asking","open-source-hedl.01d914f110":"It comes back exactly as it started.","open-source-hedl.a971507f72":"One into another","open-source-hedl.368e8081b7":"Turn","open-source-hedl.e18dd040ae":"The labels are written once, not on every line.","open-source-hedl.1c86e3fcaf":"A long list","open-source-hedl.adc4e96a47":"Sort","open-source-hedl.9b2fd08b9b":"Twice as much of it fits in one go.","open-source-hedl.63326f4aba":"A long page","open-source-hedl.852b438f91":"Read","open-source-hedl.7faa78d752":"The helper reads less to answer you.","open-source-hedl.5b63ad90f4":"A question","open-source-hedl.3562777bc6":"Ask","open-source-hedl.4efbfda1e4":"Four everyday moments","open-source-hedl.ff6e772197":"WHAT YOU NOTICE","open-source-hedl.9dd8e5cd3e":"CONVERTING","open-source-hedl.11691e308b":"SORTING","open-source-hedl.f09e6eca99":"READING","open-source-hedl.3304034a18":"ASKING","open-source-hedl.5a45057234":"You will never start HEDL or even see it. But every time one of those AI helpers is doing a job for you, it is quietly there, making the helper cheaper to run. Pick a moment and see what it is doing behind the curtain.","open-source-hedl.47166ee369":"in moments like these","open-source-hedl.b8155477b7":"It is working","open-source-hedl.a39c218b1c":"Where it helps you","open-source-hedl.4d8a37a4df":"Nothing lost","open-source-hedl.049a789b12":"The compact message is turned back into the full version and checked before it is used.","open-source-hedl.052ebf112c":"Is it trustworthy","open-source-hedl.4b641a986a":"Yes. The short way folds back into the long way, word for word.","open-source-hedl.ef584e50d9":"Can I get the full version back","open-source-hedl.ed7d63e9b7":"No. The meaning stays exactly the same, every time.","open-source-hedl.806f473c03":"Will it change my answer","open-source-hedl.8f743fe40c":"No. You never see it or set it up. It does its work on its own.","open-source-hedl.537cf5c718":"Is it complicated","open-source-hedl.d08f4abf86":"And why you do not need to","open-source-hedl.b348cc3eb9":"WHAT YOU MIGHT WORRY ABOUT","open-source-hedl.e72d571cea":"NO SURPRISES","open-source-hedl.090824d8fc":"THE ANSWER","open-source-hedl.eb6ce24aa7":"THE BILL","open-source-hedl.deab162f58":"Think of writing the same address on a hundred envelopes. You could write it in full every time, or write it once and keep it short on the rest. The letters still arrive. HEDL does the short way for you, and here is what that means for your side of things.","open-source-hedl.c4730a34aa":"and the short way","open-source-hedl.1f6f4fd113":"The long way,","open-source-hedl.41d7edc25a":"What it does for you","open-source-hedl.3da6958c6c":"Fewer words","open-source-hedl.3bb560bf03":"It works quietly in the background.","open-source-hedl.c80e161083":"No setup","open-source-hedl.d4c2792a72":"Hidden","open-source-hedl.c8ce028415":"Nothing is left out, nothing changes.","open-source-hedl.9525ae1781":"Safe","open-source-hedl.1490fb3d08":"Fewer words means less to pay for.","open-source-hedl.370515627b":"Smaller bill","open-source-hedl.f4cfc474fa":"Cheaper","open-source-hedl.3db7cf9747":"The same message, said in fewer words.","open-source-hedl.9409e761c8":"Same thing","open-source-hedl.8f15733bbe":"Shorter","open-source-hedl.5222f8307a":"Four things worth knowing","open-source-hedl.11405ca028":"THE SHORT VERSION","open-source-hedl.541a5acda1":"NOTHING TO LEARN","open-source-hedl.c7c0a993dc":"EVERYDAY","open-source-hedl.b5d2541ad8":"NO JARGON","open-source-hedl.5b0436e4fb":"No technical words, no charts. Just the idea, the way you would explain it to a friend. HEDL is a shorter way for computers to say the same thing to an AI helper, so the same answer costs less to get.","open-source-hedl.f572c73bb9":"why it matters to you","open-source-hedl.66e1586456":"What is this,","open-source-hedl.704869f20f":"In plain words","open-source-hedl.dcf7ee79d7":"Designed and maintained in Europe","open-source-hedl.44ab85a252":"Origin","open-source-hedl.2c1590712d":"Local, no external service","open-source-hedl.81280a4cd9":"+10.3pp vs JSON average","open-source-hedl.7974ed39d3":"56 percent fewer vs JSON","open-source-hedl.ad6c31bfe9":"Token savings","open-source-hedl.0bc6244663":"Remove the proxy; JSON continues","open-source-hedl.1ef1565eea":"Exit path","open-source-hedl.3123868410":"Placement","open-source-hedl.93ffb9d5ca":"Library · MCP · WASM","open-source-hedl.cba0624eb5":"Procurement sheet","open-source-hedl.aba3dee917":"What it saves","open-source-hedl.bb2c43e3b4":"HEDL sits in front of your AI tools and sends the same requests in 56 percent fewer tokens while speaking plain JSON to the rest of your stack. Nothing downstream changes: the result is a lower token bill, an accuracy difference measured in one benchmark run, and conversion that can run on your own machines without moving data to another service.","open-source-hedl.98dfb8636a":"keep the same data","open-source-hedl.5ae661da05":"HEDL","open-source-hedl.13446fdcac":"No, the round trip is checked","open-source-hedl.b098341409":"Can anything go missing","open-source-hedl.f7d0c776f7":"Designed and maintained in Europe","open-source-hedl.d4ea241daf":"Who made it","open-source-hedl.eb1b344997":"No, it works in the background","open-source-hedl.65ba3c66e0":"Do you install it","open-source-hedl.4ced6f796c":"Yes, turned back word for word","open-source-hedl.a6ee54b76c":"Can it be undone","open-source-hedl.2840eb79f5":"No, the meaning stays the same","open-source-hedl.6ca8aeda09":"Does it change answers","open-source-hedl.1ab5fd198f":"Fewer words means a smaller bill","open-source-hedl.750c2096f3":"Send less, pay less","open-source-hedl.bed955aea1":"A shorter way to say the same thing","open-source-hedl.f8c94a3b4b":"What it is","open-source-hedl.afae86e6ff":"See how it works","open-source-hedl.3623133c54":"You do not need to know any of the technical words to follow this. When a computer talks to one of those AI helpers, it sends a lot of words, and every word is paid for, a little like a phone bill counted by the word. HEDL is a way of saying exactly the same thing with far fewer words. Nothing is left out, the meaning stays the same, and the answer you get back does not change. It just costs less, and you never have to touch it.","open-source-hedl.4176a743d5":"said in fewer words","open-source-hedl.ff5bf20b56":"The same answer,","open-source-hedl.3d89b769e2":"MCP server, library, WASM","open-source-hedl.6cca1b6ffd":"571 tasks, 7 datasets","open-source-hedl.b930a90d78":"Benchmark scope","open-source-hedl.73d6d47b34":"56% vs JSON (measured)","open-source-hedl.a05285ddb3":"Rust (MSRV 1.70)","open-source-hedl.89b86ab0e6":"Language","open-source-hedl.942587e61e":"Project sheet","open-source-hedl.d4a74b503a":"Read the docs","open-source-hedl.e105fa2b3e":"Every API call to a large language model is priced by the token. Most of that token budget is not data; it is overhead: repeated field names, nested brackets, quote marks around every string. HEDL sent exactly the same information using 56 percent fewer tokens in one benchmark run across 571 extraction tasks and 7 datasets, with a +10.3pp accuracy difference from JSON in that run. Drop in the MCP server and your existing JSON pipeline keeps working exactly as before.","open-source-hedl.f6a06b5343":"Measured","open-source-hedl.1725f8d195":"More meaning per token","open-source-hedl.1dabba21cd":"✓","open-source-hedl.f3ec8e880a":"EDGE","open-source-hedl.b6cac10f50":"EXIT","open-source-hedl.54d0252c76":"Auditable","open-source-hedl.a4ae750b61":"Anchor","open-source-hedl.d3ae14a50b":"Why","open-source-hedl.b073f6c68e":"Scope","open-source-hedl.4de79687dc":"hedl-mcp","open-source-hedl.be8c03b464":"hedl-cli","open-source-hedl.1c076dfe4e":"hedl-wasm","open-source-hedl.00db09ab06":"PLAIN","open-source-hedl.264fbe323a":"EVERYDAY","open-source-hedl.ddd2c0300d":"CALM","open-source-hedl.4acc8a696c":"PLACEMENT","open-source-hedl.0f4b423c4b":"MEASURED","open-source-hedl.6e10953f3e":"PATH","open-source-hedl.db1a21a0bc":"JSON","open-source-hedl-rebuild.hero.titleLead":"HEDL. A token-efficient data format","open-source-hedl-rebuild.hero.titleAccent":"for LLMs","open-source-hedl-rebuild.hero.eyebrow":"OPEN SOURCE / DATA FORMAT","open-source-hedl-rebuild.hero.ledger":"HEDL / FORMAT","open-source-hedl-rebuild.hero.pill":"APPLICATION/HEDL","open-source-hedl-rebuild.hero.row0":"Licence","open-source-hedl-rebuild.hero.row1":"MSRV","open-source-hedl-rebuild.hero.row2":"Media type","open-source-hedl-rebuild.hero.row3":"500 flat records","open-source-hedl-rebuild.breadcrumb.open":"Open Source","open-source-hedl-rebuild.close.cta":"Get started","open-source-hedl-rebuild.close.headline":"Half the words","open-source-hedl-rebuild.next.eyebrow":"Continue exploring","open-source-hedl-rebuild.hero.technical.lede":"HEDL (Hierarchical Entity Data Language) is an open-source, token-efficient text data format under Apache 2.0 for LLMs and structured records. Its Rust reader, writer, validators and servers keep the format inspectable and reversible.","open-source-hedl-rebuild.hero.business.lede":"HEDL (Hierarchical Entity Data Language) is an open-source, token-efficient table format under Apache 2.0 for LLM requests. Start locally with cargo add hedl; repeated labels disappear while conversion stays local and reversible. In one measured run, 500 flat records were 48.7% smaller than minified JSON.","open-source-hedl-rebuild.hero.consumer.lede":"HEDL (Hierarchical Entity Data Language) is an open-source text format under Apache 2.0 for lists sent to LLMs. It writes the column labels once at the top, then keeps the same facts in each row.","open-source-hedl-rebuild.hero.primary":"Inspect the format","open-source-hedl-rebuild.hero.secondary":"Read the specification","open-source-hedl-rebuild.close.body":"Add the Rust library with cargo add hedl or install the CLI with cargo install hedl-cli. Run the reader locally and inspect the bytes yourself.","open-source-hedl-rebuild.close.install":"Install the CLI","open-source-hedl-rebuild.close.source":"Browse the source","open-source-hedl-rebuild.close.demo":"Try the browser demo","open-source-hedl-rebuild.close.docs":"Read the documentation","open-source-hedl-rebuild.next.reed.category":"CONTENT-ADDRESSED PARSING","open-source-hedl-rebuild.next.reed.body":"A content-addressed parsing pipeline.","open-source-hedl-rebuild.next.bitweave.category":"RETRIEVAL","open-source-hedl-rebuild.next.bitweave.body":"Sub-millisecond retrieval for richer prompts.","open-source-hedl-rebuild.next.nexus.category":"PRODUCTS","open-source-hedl-rebuild.next.nexus.body":"Multi-agent orchestration with HEDL throughout.","oss-hedl-bench-scope.e566c66130":"HEDL / BENCHMARK SCOPE","oss-hedl-bench-scope.26411c6c84":"Benchmark corpus and methodology ship in the repository.","oss-hedl-bench-scope.c79e2545a1":"more correct answers per token than JSON","oss-hedl-bench-scope.3a344d306b":"Models tested","oss-hedl-bench-scope.5932c9c2de":"real-world datasets","oss-hedl-bench-scope.fe051fed2c":"extraction questions","oss-hedl-bench-scope.bff36ebc7e":"Test scope","oss-hedl-bench-scope.19d59f3fa3":"Higher is better. HEDL returns 2.4x more correct answers per token than JSON.","oss-hedl-bench-scope.4551d61433":"Higher is better. HEDL extracts 80.4%, 10.3 points above JSON.","oss-hedl-bench-scope.30d02db8c1":"Shorter is better. HEDL carries the same answers in 56 percent fewer tokens than JSON.","oss-hedl-bench-scope.b2bb7604c8":"Metric","oss-hedl-bench-scope.c38d9aede4":"published, reproducible","oss-hedl-bench-scope.1ed0871e23":"571 questions, 7 datasets, 6 formats","oss-hedl-bench-scope.26c62ca3f5":"perK","oss-hedl-bench-scope.5a7bd4149d":"en-US","oss-hedl-bench-scope.3ddeecdff4":"Tokens per question","oss-hedl-bench-scope.e46c92d993":"Extraction accuracy","oss-hedl-bench-scope.b0e85a8a10":"Correct answers per 1K tokens","oss-hedl-bench-scope.a07cb2cc48":"NVIDIA GLM-4.7","oss-hedl-bench-scope.f38a903923":"DeepSeek Chat","oss-hedl-bench-scope.7b966ed3dc":"Mistral Large","oss-hedl-bench-scope.3281188351":"CSV","oss-hedl-bench-scope.faa8c01f2d":"TOON","oss-hedl-bench-scope.c0aa9ef764":"XML","oss-hedl-bench-scope.ffbcb202b0":"YAML","oss-hedl-bench-scope.031a4e76f0":"JSON","oss-hedl-bench-scope.d007bb1b85":"HEDL","oss-hedl-business-glyph.dd2c7eb783":"Token bill","oss-hedl-business-glyph.f31f941af9":"fewer tokens, same data","oss-hedl-business-glyph.12a3a4f498":"Formats","oss-hedl-business-glyph.02b9c65166":"JSON · YAML · XML · CSV · Parquet · TOON","oss-hedl-business-glyph.6e3aac9200":"Conversions","oss-hedl-business-glyph.296fbd571e":"12 pairs, both ways","oss-hedl-business-glyph.bf5dde2d06":"Validation","oss-hedl-business-glyph.6eef664840":"27 conformance files","oss-hedl-business-glyph.fc81e1d40e":"15 accept · 12 refuse","oss-hedl-business-glyph.b0cfc2c0ce":"On JSON today","oss-hedl-business-glyph.820559510a":"3,000 / mo","oss-hedl-business-glyph.de144d2e09":"On HEDL","oss-hedl-business-glyph.98c23c3df8":"1,320 / mo","oss-hedl-business-glyph.fef8e889dc":"Interfaces","oss-hedl-business-glyph.6b5eb5323e":"Library · MCP · WASM","oss-hedl-business-glyph.6da13addb0":"Placement","oss-hedl-business-glyph.e272d6100f":"Before the model","oss-hedl-business-glyph.c10e83c453":"Data residency","oss-hedl-business-glyph.d9aa555285":"Local, no egress","oss-hedl-business-glyph.44ab85a252":"Origin","oss-hedl-business-glyph.ae0fd596f4":"European foundation","oss-hedl-business-glyph.005fd98d0b":"Outcomes board","oss-hedl-business-glyph.d007bb1b85":"HEDL","oss-hedl-business-glyph.6cc3364a24":"MEASURED","oss-hedl-business-glyph.8ac7aa173f":"Published surface","oss-hedl-business-glyph.86c8b48d59":"Monthly token bill","oss-hedl-business-glyph.e9bd723731":"Same requests, same answers","oss-hedl-business-glyph.7d124660c4":"56% lower","oss-hedl-business-glyph.ac3686dbbe":"No re-platform","oss-hedl-business-glyph.86b6e48353":"Reversible","oss-hedl-business-glyph.818c05f658":"Data stays local","oss-hedl-business-glyph.measure.payload-label":"Payload size","oss-hedl-business-glyph.measure.payload-value":"48.7%","oss-hedl-business-glyph.measure.payload-note":"smaller than minified JSON","oss-hedl-business-glyph.measure.json-size":"35.8 KB","oss-hedl-business-glyph.measure.hedl-size":"18.4 KB","oss-hedl-business-glyph.measure.json-label":"JSON, minified","oss-hedl-business-glyph.measure.prompt-label":"Measured run","oss-hedl-business-glyph.measure.prompt-note":"571 questions · one model","oss-hedl-business-glyph.measure.comparison-label":"500 flat records, measured","oss-hedl-business-glyph.measure.comparison-foot":"same 500 flat records","oss-hedl-business-glyph.measure.comparison-saving":"48.7% smaller vs minified JSON in this run","oss-hedl-compliance-sheet.e566c66130":"HEDL / DEPLOYMENT EVIDENCE","oss-hedl-compliance-sheet.942587e61e":"Project sheet","oss-hedl-compliance-sheet.920e458269":"local, no external service","oss-hedl-compliance-sheet.fef8e889dc":"Interfaces","oss-hedl-compliance-sheet.6b5eb5323e":"Library · MCP · WASM","oss-hedl-compliance-sheet.c3d33240c4":"Tasks benchmarked","oss-hedl-compliance-sheet.ff14e25f15":"Accuracy gain","oss-hedl-compliance-sheet.d5a3cabacb":"Fewer tokens","oss-hedl-compliance-sheet.2cf7f7e5d6":"JSON, YAML, XML, CSV, Parquet","oss-hedl-compliance-sheet.9151f8433f":"Conversion","oss-hedl-compliance-sheet.f0d925b213":"Zero unsafe code","oss-hedl-compliance-sheet.9bf8837cbd":"Core parser","oss-hedl-compliance-sheet.780d8d3c08":"10,000+ across 19 crates","oss-hedl-compliance-sheet.39fdec1194":"Tests","oss-hedl-compliance-sheet.3123868410":"Repository","oss-hedl-compliance-sheet.a2bf05dbd6":"Rust, MSRV 1.70","oss-hedl-compliance-sheet.89b86ab0e6":"Language","oss-hedl-compliance-sheet.6e4ff8e97c":"Bidirectional, lossless conversion. The same data semantics, encoded more densely, restored exactly.","oss-hedl-compliance-sheet.f3b36f28ca":"Lossless","oss-hedl-compliance-sheet.4af9999f4a":"A drop-in proxy. Remove it and your stack is back on plain JSON, with no migration to undo.","oss-hedl-compliance-sheet.86b6e48353":"Reversible","oss-hedl-compliance-sheet.2721c6f194":"The Rust core has zero dependencies and no unsafe code; shipped tools keep their runtime dependencies explicit.","oss-hedl-compliance-sheet.fb85566784":"Supply chain","oss-hedl-compliance-sheet.419c168a61":"No vendor sits in the path. Conversion runs locally, so records never leave your own infrastructure.","oss-hedl-compliance-sheet.c10e83c453":"Data residency","oss-hedl-consumer-glyph.9d83e0615e":"It works quietly in the background","oss-hedl-consumer-glyph.10fc3853d2":"Nothing to set up","oss-hedl-consumer-glyph.e238524217":"Both letters reach the same place","oss-hedl-consumer-glyph.10fac79d24":"Same answer arrives","oss-hedl-consumer-glyph.3da6958c6c":"Fewer words","oss-hedl-consumer-glyph.e5f07af9b4":"The short way","oss-hedl-consumer-glyph.77cd733ff2":"folded shorter","oss-hedl-consumer-glyph.9e9ff4a7b1":"More words","oss-hedl-consumer-glyph.94f72376d3":"The long way","oss-hedl-consumer-glyph.d007bb1b85":"HEDL","oss-hedl-consumer-glyph.c5ab118f7b":"The same thing, said in fewer words","oss-hedl-consumer-glyph.c9684fd966":"You never have to touch it","oss-hedl-consumer-glyph.64ce413ab5":"Nothing is left out","oss-hedl-consumer-glyph.5949143c57":"The same answer comes back","oss-hedl-consumer-glyph.8bcd73bb9f":"today, plain words.","oss-hedl-consumer-glyph.4bbe75e2cb":"Lyon weather,","oss-hedl-consumer-glyph.aba7b1748b":"thank you very much.","oss-hedl-consumer-glyph.9c64ab7347":"of Lyon, in plain words,","oss-hedl-consumer-glyph.41b34e9814":"for today, in the city","oss-hedl-consumer-glyph.2f5d4f9be4":"Please send the weather","oss-hedl-cost-meter.60cf2bf910":"A small service","oss-hedl-cost-meter.e9ddd9aa6c":"A handful of AI features in steady use.","oss-hedl-cost-meter.bcbfb3d54e":"A busy product","oss-hedl-cost-meter.f1b2e28fe2":"The README reference point.","oss-hedl-cost-meter.d037453a9f":"A heavy pipeline","oss-hedl-cost-meter.189611c4ab":"Bulk extraction billed by volume.","oss-hedl-cost-meter.5a7bd4149d":"en-US","oss-hedl-cost-meter.c5e74f063a":"same data, 56 percent fewer tokens","oss-hedl-cost-meter.4699af87ef":"Monthly volume","oss-hedl-cost-meter.b51366690f":"tokens / month","oss-hedl-cost-meter.29844bc989":"JSON today","oss-hedl-cost-meter.5b4056d61d":"With HEDL","oss-hedl-cost-meter.c50179005a":"saved per month","oss-hedl-cost-meter.83101a3ef7":"saved per year","oss-hedl-cost-meter.b76aee974c":"Same","oss-hedl-cost-meter.24da29c358":"data, same answers","oss-hedl-cost-meter.e2c58504c9":"Figures follow the published reference: a service at 1B tokens a month saves $1,680 a month.","oss-hedl-cost-meter.f40060b6a9":"HEDL / COST PER TOKEN","oss-hedl-everyday-scenes.3e840c3f22":"You never start any of this. It happens on its own, every time a helper is at work,\n          whichever of these you are doing.","oss-hedl-everyday-scenes.d92a1780f4":"HEDL does this","oss-hedl-everyday-scenes.b827d6e115":"You do this","oss-hedl-everyday-scenes.34639fccaa":"Everyday moments","oss-hedl-everyday-scenes.f360f2e843":"Comes back unchanged","oss-hedl-everyday-scenes.1d91cb1a0a":"HEDL only changes the shape of the message for the trip there and back. What you started with comes back exactly the same, with nothing added and nothing lost.","oss-hedl-everyday-scenes.89e162521d":"You ask the helper to take something in one shape and hand it back in another, neat and ready to use.","oss-hedl-everyday-scenes.256bc5ffc9":"Turning one thing into another","oss-hedl-everyday-scenes.4d39d96fbd":"Labels written once","oss-hedl-everyday-scenes.7f29e666e9":"A plain list repeats the same little labels on every single line. HEDL writes the labels once at the top, then only the entries, so the list takes far fewer words to send.","oss-hedl-everyday-scenes.5b72026707":"You ask the helper to go down a long list, a list of names, orders, or appointments, and tidy it up.","oss-hedl-everyday-scenes.6d4b0f3161":"Sorting through a long list","oss-hedl-everyday-scenes.da2e7a54b5":"Twice as much fits","oss-hedl-everyday-scenes.e62680c425":"A long document is a lot of words to send. Folded the HEDL way, twice as much of it fits in one go, so the helper sees more of the page at once.","oss-hedl-everyday-scenes.847271fc1b":"You hand the helper a long letter or report and ask it to pull out the parts you care about.","oss-hedl-everyday-scenes.ee0f33551a":"Working through a long document","oss-hedl-everyday-scenes.302a58506b":"Same answer, sooner paid for","oss-hedl-everyday-scenes.873db7c6da":"Behind the curtain, the helper had to read a lot of background first. HEDL folds that background up small, so it reads the same thing with far less to chew through.","oss-hedl-everyday-scenes.9351423803":"You type a question into one of those AI helpers and it writes back an answer in a moment.","oss-hedl-everyday-scenes.64b3c4f4bd":"Asking a helper a question","oss-hedl-everyday-wins.d7ff90c706":"Same answer, said in fewer words. That is all that changes.","oss-hedl-everyday-wins.e5f07af9b4":"The short way","oss-hedl-everyday-wins.94f72376d3":"The long way","oss-hedl-everyday-wins.bc11b12738":"You can turn the short way back into the long way at any time, word for word.","oss-hedl-everyday-wins.2c1710b70e":"You wonder whether you can ever get the full version back if you need it.","oss-hedl-everyday-wins.69bfc818a3":"No surprises","oss-hedl-everyday-wins.1aede2a56d":"Nothing is left out. The shorter way carries exactly the same meaning as the long way.","oss-hedl-everyday-wins.bbbdec0943":"You worry that saying it shorter might leave something out or change the meaning.","oss-hedl-everyday-wins.7011b1a621":"The answer","oss-hedl-everyday-wins.4ee872eb8e":"The same message, said with far fewer words, so the bill at the end is smaller.","oss-hedl-everyday-wins.c67f9b221d":"Every word the helper reads is paid for, even the words it repeats over and over.","oss-hedl-everyday-wins.33620369b5":"The bill","oss-hedl-format-lab.e566c66130":"HEDL / FORMAT LAB","oss-hedl-format-lab.9bb8b90ae6":"fewer tokens","oss-hedl-format-lab.5a7bd4149d":"en-US","oss-hedl-format-lab.d007bb1b85":"HEDL","oss-hedl-format-lab.031a4e76f0":"JSON","oss-hedl-format-lab.f6ba279f17":"Tokens for the same payload","oss-hedl-format-lab.32ff4af810":"Schema declared once. Rows are pure positional data.","oss-hedl-format-lab.09086e805c":"id, name and role are paid on every record.","oss-hedl-format-lab.acfd08d506":"lossless round trip","oss-hedl-format-lab.93a7f22476":"Datasets","oss-hedl-format-lab.dcd72ea8d2":"Config files","oss-hedl-format-lab.4ba7c734ba":"GHI-789,Gizmo,","oss-hedl-format-lab.ee2bab214f":"DEF-456,Gadget,","oss-hedl-format-lab.d2d73a120b":"ABC-123,Widget,","oss-hedl-format-lab.546e7e8f54":"sku, name, price","oss-hedl-format-lab.b6a98918b0":"%S:Product:","oss-hedl-format-lab.47306018f1":"\"Gizmo\"","oss-hedl-format-lab.a3b749d4af":"\"Gadget\"","oss-hedl-format-lab.e9430a3701":"\"Widget\"","oss-hedl-format-lab.6fd4a15b27":"Product catalog","oss-hedl-format-lab.ba19d53a53":"u3,Carol,","oss-hedl-format-lab.417b42ca5a":"u2,Bob,","oss-hedl-format-lab.b35f1f0cb1":"u1,Alice,","oss-hedl-format-lab.7dd039aed4":"id, name, role","oss-hedl-format-lab.8998ca27c2":"%S:User:","oss-hedl-format-lab.1c709bac84":"\"Carol\"","oss-hedl-format-lab.b6b883bfeb":"\"Bob\"","oss-hedl-format-lab.ab209bc5ec":"\"Alice\"","oss-hedl-format-lab.40e6410106":"User records","oss-hedl-hero-glyph.6cca1b6ffd":"571 tasks, 7 datasets","oss-hedl-hero-glyph.7a5097d6bc":"15,697 → 6,912 tokens","oss-hedl-hero-glyph.f7cc2d2eec":"HEDL 2.0","oss-hedl-hero-glyph.8475cb8ef1":"to-json users.hedl","oss-hedl-hero-glyph.237ad8c58f":"strings unquoted","oss-hedl-hero-glyph.93d61dbf3c":"no brackets","oss-hedl-hero-glyph.5eef6d7a86":"no repeated keys","oss-hedl-hero-glyph.9bb8b90ae6":"fewer tokens","oss-hedl-hero-glyph.c0ae8f6ea8":"Saved","oss-hedl-hero-glyph.dc614b1139":"Schema, User","oss-hedl-hero-glyph.2f24c67ee5":"users.hedl, no issues found","oss-hedl-hero-glyph.1ae384be52":"hedl lint users.hedl","oss-hedl-hero-glyph.c38c6c1f3a":"Tokens","oss-hedl-hero-glyph.041a5dec48":"Format","oss-hedl-hero-glyph.a380d50627":"Estimated Tokens (LLM context)","oss-hedl-hero-glyph.a1876da04e":"hedl stats users.hedl","oss-hedl-hero-glyph.b5dab9598c":"u2,Bob,bob@co,user","oss-hedl-hero-glyph.706830b37d":"u1,Alice,alice@co,admin","oss-hedl-hero-glyph.84e1290c80":"hedl out","oss-hedl-hero-glyph.b6b883bfeb":"\"Bob\"","oss-hedl-hero-glyph.ab209bc5ec":"\"Alice\"","oss-hedl-hero-glyph.8dcaa31dfb":"json in","oss-hedl-hero-glyph.0d2b471e45":"hedl from-json users.json","oss-hedl-hero-glyph.2805885e0f":"schema declared once","oss-hedl-hero-glyph.ee00fe30ef":", Hierarchical Entity Data Language","oss-hedl-hero-glyph.f3b36f28ca":"Lossless","oss-hedl-hero-glyph.cc73ae0f54":"json → hedl → json","oss-hedl-hero-glyph.7419737147":"HEDL [","oss-hedl-hero-glyph.cb67aba45e":"XML (minified)","oss-hedl-hero-glyph.ffbcb202b0":"YAML","oss-hedl-hero-glyph.776cfa3c8e":"JSON (pretty)","oss-hedl-hero-glyph.b537be17fa":"JSON (minified)","oss-hedl-hero-glyph.d007bb1b85":"HEDL","oss-hedl-hero-glyph.de4f4b5446":"User:[id, name, email, role]","oss-hedl-hero-glyph.measure.caption":"Measured file size · 500 flat records","oss-hedl-hero-glyph.measure.size":"Bytes","oss-hedl-hero-glyph.measure.comparison":"HEDL saving","oss-hedl-hero-glyph.measure.result":"result","oss-hedl-hero-glyph.measure.saving-minified":"48.7%","oss-hedl-hero-glyph.measure.saving-indented":"67.7%","oss-hedl-hero-glyph.measure.saving":"smaller than minified JSON","oss-hedl-hero-glyph.measure.range":"35.8 KB → 18.4 KB","oss-hedl-hero-glyph.measure.summary":"500 flat records","oss-hedl-integration-console.e566c66130":"HEDL / INTEGRATION PATH","oss-hedl-integration-console.6b5eb5323e":"Start at one busy interface","oss-hedl-integration-console.4fd7b25687":"Deployment targets","oss-hedl-integration-console.147c4802cb":"no app change","oss-hedl-integration-console.054b06cc7b":"hedl ~","oss-hedl-integration-console.0fabc72447":"Remove the proxy","oss-hedl-integration-console.d74197fdde":"Roll back","oss-hedl-integration-console.b52acaf260":"Tokens logged each call","oss-hedl-integration-console.d4fd0a232e":"Measure","oss-hedl-integration-console.8879ce9e85":"Agent talks HEDL","oss-hedl-integration-console.136e9d668e":"Front the tool","oss-hedl-integration-console.1bcb5e9171":"One binary or crate","oss-hedl-integration-console.fd6c3ebf7b":"Install","oss-hedl-integration-console.ca1b66cfd4":"converts client side, nothing leaves the browser","oss-hedl-integration-console.cff3c4e78f":"await init();","oss-hedl-integration-console.ca0bb28ff9":"import init, { fromJson } from \"hedl-wasm\";","oss-hedl-integration-console.e22a350436":"npm install hedl-wasm","oss-hedl-integration-console.68aa0310d3":"Run HEDL in the browser or at the edge. The same engine compiled to WebAssembly for client-side conversion.","oss-hedl-integration-console.b3d5f99d67":"zero unsafe in the core parser","oss-hedl-integration-console.35e39da7b1":"let json = to_json(&doc)?;","oss-hedl-integration-console.62885c9bbd":"let doc  = parse(src)?;","oss-hedl-integration-console.571222078d":"use hedl::{parse, to_json, from_json};","oss-hedl-integration-console.5082488cc1":"cargo add hedl","oss-hedl-integration-console.77f14cb52a":"Embed the engine directly in your service. Parse, validate and convert in process, no external server in the path.","oss-hedl-integration-console.2ead852dbf":"ok   lossless, run hedl to-json to reverse","oss-hedl-integration-console.089a5e26f3":" |u1,Alice,admin","oss-hedl-integration-console.cebee57ec1":"%S:User:[id,name,role]","oss-hedl-integration-console.4afd5164da":"echo '{\"users\":[...]}' | hedl from-json","oss-hedl-integration-console.893cb8c284":"cargo install hedl-cli","oss-hedl-integration-console.f736c457bf":"Convert, validate, lint and format from the terminal. Round-trip JSON to HEDL and back on the command line.","oss-hedl-integration-console.ec8e85b099":"response  hedl  ->  json   restored, lossless","oss-hedl-integration-console.3c5b049a7d":"request   json 15,697 tok  ->  hedl 6,912 tok   -56%","oss-hedl-integration-console.20a47e334c":"hedl-mcp listening, auto-convert on","oss-hedl-integration-console.bae2ffb5e1":"\"hedl\": { \"command\": \"hedl-mcp\", \"args\": [\"--auto-convert\"] }","oss-hedl-integration-console.ff48528ff4":"# .mcp config, add one block","oss-hedl-integration-console.44765133c0":"Drop in front of any MCP tool. The agent talks HEDL, the tool keeps speaking JSON. Zero application code change.","oss-hedl-interop-board.031a4e76f0":"JSON","oss-hedl-interop-board.5241dda61d":"REST APIs and web services","oss-hedl-interop-board.ffbcb202b0":"YAML","oss-hedl-interop-board.4aae675999":"Kubernetes and CI/CD configs","oss-hedl-interop-board.c0aa9ef764":"XML","oss-hedl-interop-board.dd2f3247e5":"Enterprise systems and SOAP","oss-hedl-interop-board.3281188351":"CSV","oss-hedl-interop-board.a52e2f5949":"Spreadsheets and data analysis","oss-hedl-interop-board.b6d723c442":"Parquet","oss-hedl-interop-board.64eb1972f0":"Data lakes and analytics","oss-hedl-interop-board.1b6c09047b":"Neo4j Cypher","oss-hedl-interop-board.cfed0f8134":"Graph databases","oss-hedl-interop-board.faa8c01f2d":"TOON","oss-hedl-interop-board.8501963a5a":"Other token-efficient format","oss-hedl-interop-board.1e41ef2a5f":"No re-platforming","oss-hedl-interop-board.59e781b598":"It sits in front of your AI tools and converts at the edge. Your existing systems keep sending and receiving the formats they already use.","oss-hedl-interop-board.7dcbb222b1":"Nothing downstream changes","oss-hedl-interop-board.4dec2871df":"Your databases, reports, and spreadsheets see the same shapes as before. The denser format only exists on the trip to and from the model.","oss-hedl-interop-board.ea5025520f":"Clean exit any day","oss-hedl-interop-board.22c81493f2":"Seven formats enter and leave through the same lossless bridge; the compact representation exists only around the model call.","oss-hedl-interop-board.1657e0349d":"imports and exports both ways","oss-hedl-interop-board.b3bde5d151":"Converts in both directions","oss-hedl-interop-board.d007bb1b85":"HEDL","oss-hedl-interop-board.6f94d00923":"converts at the edge","oss-hedl-interop-board.154400cf2b":"Your AI tools","oss-hedl-interop-board.02dc44d7b0":"The model receives the dense version. Your stack keeps speaking the formats it always did.","oss-hedl-interop-board.7845f5bcf8":"56 percent fewer tokens on the trip","oss-hedl-interop-board.7cc1bde74b":"One drop-in converter sits between your systems and the model.","oss-hedl-interop-board.e566c66130":"HEDL / LOSSLESS BRIDGE","oss-hedl-overhead-anatomy.e566c66130":"HEDL / OVERHEAD ANATOMY","oss-hedl-overhead-anatomy.3868c469ee":"saved per month at one billion tokens by switching JSON to HEDL.","oss-hedl-overhead-anatomy.c2136e2efa":"wasted per API call on repeated keys for a 10,000 user dataset.","oss-hedl-overhead-anatomy.393f016588":"Schema once. Then only the data, comma separated.","oss-hedl-overhead-anatomy.d16692536e":"|u1,Alice,admin","oss-hedl-overhead-anatomy.cebee57ec1":"%S:User:[id,name,role]","oss-hedl-overhead-anatomy.d007bb1b85":"HEDL","oss-hedl-overhead-anatomy.a8b52ebbbb":"Now multiply the structural bytes by ten thousand records.","oss-hedl-overhead-anatomy.b14c81ef17":"x 1","oss-hedl-overhead-anatomy.92cd410b35":"One record, every byte tagged","oss-hedl-overhead-anatomy.31989deedb":"hover a part to isolate it","oss-hedl-overhead-anatomy.ab209bc5ec":"\"Alice\"","oss-hedl-overhead-anatomy.4cf6494a14":"The values you actually meant to send. In HEDL, this is nearly all that remains.","oss-hedl-overhead-anatomy.48cd18dabc":"Your data","oss-hedl-overhead-anatomy.01c379afb8":"A quote pair wrapped around every string, even the unambiguous ones.","oss-hedl-overhead-anatomy.7b0257ce75":"Quotes","oss-hedl-overhead-anatomy.17cb7a1df5":"Object braces and array brackets that mark structure the schema already knows.","oss-hedl-overhead-anatomy.19ab36d21e":"Brackets","oss-hedl-overhead-anatomy.2316665756":"The same field names, \"id\", \"name\", \"role\", paid again on every single record.","oss-hedl-overhead-anatomy.9bc20846df":"Repeated keys","oss-hedl-ownership.fef8e889dc":"Placement","oss-hedl-ownership.6b5eb5323e":"Before the model","oss-hedl-ownership.b6e77c24ce":"Interfaces","oss-hedl-ownership.6b0c36de27":"Language","oss-hedl-ownership.e2ae20d9ae":"Library · MCP · WASM","oss-hedl-ownership.c2098536e0":"Rust, MSRV 1.70","oss-hedl-ownership.86b6e48353":"Reversible","oss-hedl-ownership.052e38b85c":"Remove and you are back on JSON","oss-hedl-ownership.870c4d024a":"Clean insertion","oss-hedl-ownership.9ff779ba2e":"One compact layer sits before the model while the rest of your stack continues speaking JSON.","oss-hedl-ownership.3c1fdfbcf8":"Nothing to rip out later","oss-hedl-ownership.5d973734df":"It speaks HEDL to the model and plain JSON back to your own systems, so the rest of your stack does not change. Take it out and you are exactly where you started.","oss-hedl-ownership.e5b0b6355d":"Runs where you do","oss-hedl-ownership.6f5c5f5551":"Conversion can run on your own machines, in an EU region, or fully offline without an external service in the path.","oss-hedl-ownership.854bde3b23":"Measured behaviour","oss-hedl-ownership.dfa1001024":"Token reduction, answer accuracy, and lossless round trips are measured across real datasets before rollout.","oss-hedl-ownership.70723e7625":"REVERSIBLE PATH","oss-hedl-ownership.15a8cde422":"Insert it before one workflow, compare the result, then expand only where it pays.","oss-hedl-ownership.217fc9a42f":"REMOVE IT CLEANLY","oss-hedl-peace-of-mind.25d03595bd":"HEDL is one small piece that makes AI helpers cheaper to run, without changing what they\n            say back to you. You do not install it, you do not set it up, and you do not have to\n            understand it. Here is the short version of what it promises.","oss-hedl-peace-of-mind.535429d926":"You can stop worrying about this part.","oss-hedl-peace-of-mind.c8be74ffd6":"Because it says the same thing in fewer words, the helpers that use it cost less to run.","oss-hedl-peace-of-mind.f706e5551a":"A smaller bill","oss-hedl-peace-of-mind.e3d7d078fe":"Every compact message is restored and checked before the helper uses it, so nothing disappears on the way.","oss-hedl-peace-of-mind.67adde5015":"Nothing goes missing","oss-hedl-peace-of-mind.6d5b0b9d9e":"It only changes how the message is folded for sending. The meaning is kept exactly the same.","oss-hedl-peace-of-mind.daeb965ec5":"Your words stay your words","oss-hedl-peace-of-mind.192b52369f":"You never see HEDL or touch it. It sits behind the AI helper and does its work on its own.","oss-hedl-peace-of-mind.bbbf622b20":"Nothing to learn","oss-hedl-plain-answers.0f3e55e5c4":"If you are still unsure, that is fine. You never have to touch HEDL for it to do its job.","oss-hedl-plain-answers.57b117c9fe":"Tap a question to read the plain answer. There are no wrong ones to ask.","oss-hedl-plain-answers.09ae205c8f":"The questions people ask first","oss-hedl-plain-answers.2b0f17506d":"The helpers you use cost money to run, counted a little like a phone bill by the word. Saying the same thing in fewer words means a smaller bill, which keeps those helpers cheaper to offer.","oss-hedl-plain-answers.91f813b073":"So what is the point of it for me?","oss-hedl-plain-answers.6abbc02007":"Every compact message is restored and checked before it is used. If the full version does not come back exactly, the conversion fails instead of quietly dropping part of it.","oss-hedl-plain-answers.108164d4e6":"How do I know nothing goes missing?","oss-hedl-plain-answers.e4068b2e43":"Yes. The short way folds back into the long way whenever it needs to, word for word, with nothing added and nothing lost. It is like folding a letter to post it, then unfolding it to read.","oss-hedl-plain-answers.1b05b1a664":"Can the full version always be brought back?","oss-hedl-plain-answers.06b9f4a8eb":"No. It only changes how the message is folded up for the trip to the helper and back. The meaning is kept exactly the same, so the answer you read is the answer you would have got anyway.","oss-hedl-plain-answers.4fd0781270":"Could it change the answer I get back?","oss-hedl-plain-answers.b4be88daf8":"No. There is nothing for you to download, switch on, or learn. HEDL sits behind the AI helpers other people build, and it does its work on its own. You will never see it.","oss-hedl-plain-answers.770cdab5d1":"Do I have to install or set up anything?","oss-hedl-plain-answers.f14ac706dc":"Nothing to do","oss-hedl-plain-answers.c050628b65":"Same answer","oss-hedl-plain-answers.51183e7f1e":"Always reversible","oss-hedl-plain-answers.d274190d73":"Round trip checked","oss-hedl-plain-answers.055833b3a0":"A smaller bill","oss-hedl-plain-story.b52fd011bc":"You do not have to install anything or understand any of this. It works quietly in the\n          background, so the helpers you use cost less and stay just as clear.","oss-hedl-plain-story.f78469e2e7":"Imagine you write the same address on a hundred envelopes. You could write the full\n          address every time. Or you could write it once, then just the house number on the rest.\n          The letters still arrive. You simply did less writing. That is the whole idea behind HEDL.","oss-hedl-plain-story.1f770d7c8d":"You can change the short way back into the long way whenever you want, word for word, with nothing changed. Like folding a letter to fit an envelope, then unfolding it later to read.","oss-hedl-plain-story.c6455b9784":"It can always be turned back","oss-hedl-plain-story.eaa52e94ee":"Every word sent to an AI helper is paid for, a little like a phone bill counted by the word. Saying the same thing in fewer words means a smaller bill, with the same answer at the end.","oss-hedl-plain-story.462080683d":"Talking costs money","oss-hedl-plain-story.d37db45555":"When a computer talks to one of these AI helpers, it sends a lot of words. HEDL is a way of saying exactly the same thing with far fewer words. Nothing is left out, it is just said more plainly.","oss-hedl-plain-story.bed955aea1":"A shorter way to say the same thing","oss-hedl-rollout-plan.676c601cc8":"Try it on one workflow","oss-hedl-rollout-plan.dfa8874b40":"Your team places the converter in front of a single AI workflow. Nothing else in the stack is touched, and the rest of the business carries on exactly as before.","oss-hedl-rollout-plan.404ff8cfdb":"One workflow live, the rest untouched.","oss-hedl-rollout-plan.b172bb4450":"See the before and after","oss-hedl-rollout-plan.213e14f769":"The converter records the token count before and after on every request, so you get a plain side-by-side of what you used to send and what you send now.","oss-hedl-rollout-plan.0b97886b72":"Real token figures, not estimates.","oss-hedl-rollout-plan.288006d466":"Roll it out where it pays","oss-hedl-rollout-plan.29d6af5389":"Once the saving is proven on the pilot, the same drop-in sits in front of the rest of your AI workflows. The saving repeats wherever tokens are billed by volume.","oss-hedl-rollout-plan.ecf7526c8f":"Lower bill across every workflow.","oss-hedl-rollout-plan.9e7ae61a18":"Keep it, or walk away","oss-hedl-rollout-plan.75fadb964d":"If it ever stops paying off, remove it and you are back on plain JSON.","oss-hedl-rollout-plan.3e2b486154":"No lock-in, clean exit any day.","oss-hedl-rollout-plan.9079ec1ed4":"PILOT → MEASURE → EXPAND → EXIT","oss-hedl-rollout-plan.dc416e1088":"Step","oss-hedl-rollout-plan.2bf91cce61":"You end with","oss-hedl-rollout-plan.97b274df11":"Start on one workflow, prove it, then decide. No big-bang migration.","oss-hedl-rollout-plan.217fc9a42f":"Start at one busy interface","oss-hedl-savings-case.e729d48a29":"Same data in, same answers out. The only thing that changes is the number on the invoice.","oss-hedl-savings-case.a6c1933f89":"What it de-risks","oss-hedl-savings-case.aba3dee917":"What it saves","oss-hedl-savings-case.0e4b5a2614":"The round trip is lossless, so the same data comes back exactly as it went in.","oss-hedl-savings-case.b5790a6dce":"No data surprises","oss-hedl-savings-case.a2af56ebba":"Remove the proxy and you are back on plain JSON, byte for byte, with no data migration to undo.","oss-hedl-savings-case.cbd17d59ba":"Clean exit","oss-hedl-savings-case.cc885c1e4f":"It sits in front of your tools and speaks plain JSON to the rest of your stack. Nothing downstream changes.","oss-hedl-savings-case.1e41ef2a5f":"No re-platforming","oss-hedl-savings-case.81535bdf88":"2.4x answers per token","oss-hedl-savings-case.b1eac55c48":"For every thousand tokens you pay for, you get more usable answers back than with JSON.","oss-hedl-savings-case.2fd3f0e442":"More work per spend","oss-hedl-savings-case.de1332a0d9":"+10.3pp accuracy","oss-hedl-savings-case.597626951e":"The denser format is also clearer to the model, so extraction is more accurate, not less.","oss-hedl-savings-case.a39bb45317":"Better answers","oss-hedl-savings-case.e19a805469":"56 percent fewer tokens","oss-hedl-savings-case.95034211b9":"The same request reaches the model in fewer tokens, and tokens are what you are billed for.","oss-hedl-savings-case.82b487b068":"Lower token bill","oss-hedl-sovereignty.01063afd00":", no vendor in the path, no border crossed without your say.","oss-hedl-sovereignty.401113aa26":"Local conversion means no data leaves your infrastructure. Run it on-premise, in an\n          EU region, or fully offline. The format never phones home.","oss-hedl-sovereignty.81bfd6ed8b":"Runs on your own ground","oss-hedl-sovereignty.5a50843797":"Designed and maintained in Europe, with no hosted service required in the conversion path.","oss-hedl-sovereignty.e5b0b6355d":"European engineering","oss-hedl-sovereignty.cd01695dd9":"It does not call out to anyone. Nothing crosses a border you did not approve, which is the whole point of running it yourself.","oss-hedl-sovereignty.586f76b845":"No foreign dependency","oss-hedl-sovereignty.c8dca1632d":"There is no external service in the path. The conversion runs on your own machines, so your data never has to leave your infrastructure to be made denser.","oss-hedl-sovereignty.23c54e04c4":"Data stays where you run it","oss-hedl-value-ledger.e566c66130":"HEDL / VALUE LEDGER","oss-hedl-value-ledger.148f471acc":"No external service. Conversion runs on your own machines.","oss-hedl-value-ledger.e729d48a29":"Same data in, same answers out. The only thing that changes is the number on the invoice.","oss-hedl-value-ledger.aba3dee917":"What it saves","oss-hedl-value-ledger.9b0a1d362a":"same data, lower bill","oss-hedl-value-ledger.dcf7a637b8":"$1,680 saved","oss-hedl-value-ledger.009ee14061":"At 1B tokens/month","oss-hedl-value-ledger.01a82a3b27":"56 percent fewer","oss-hedl-value-ledger.799b496b24":"Reduction","oss-hedl-value-ledger.bf2a711acd":"6,912 / question","oss-hedl-value-ledger.0b06f4832f":"Tokens, HEDL","oss-hedl-value-ledger.28e0196f03":"15,697 / question","oss-hedl-value-ledger.9fe04e84c4":"Tokens, JSON","oss-hedl-value-ledger.a05a31c3d2":"Lower monthly bill","oss-hedl-value-ledger.72f181c1c3":"Per-row token cost drops with the format, so large recurring jobs cost noticeably less to run.","oss-hedl-value-ledger.37dc9563ce":"Bulk processing billed by volume.","oss-hedl-value-ledger.65c8941261":"Batch jobs","oss-hedl-value-ledger.de1332a0d9":"+10.3pp accuracy","oss-hedl-value-ledger.830831db44":"Fewer tokens spent on syntax, more on the answer. Measured +10.3 accuracy points over JSON.","oss-hedl-value-ledger.e2b3f1ee3a":"Pulling structured fields out of records.","oss-hedl-value-ledger.c97121dd66":"Document extraction","oss-hedl-value-ledger.80ad3c43fc":"2x the context","oss-hedl-value-ledger.20ce5ee97e":"The same context window holds twice the material, so the model sees more of what matters.","oss-hedl-value-ledger.6bb761c300":"Apps that feed documents to the model as context.","oss-hedl-value-ledger.b76f297789":"Retrieval (RAG)","oss-hedl-value-ledger.f53bf5cf39":"Compounds per call","oss-hedl-value-ledger.d58b2f6b3f":"Every tool result comes back denser, so the saving stacks up across a long chain of calls.","oss-hedl-value-ledger.a8bc3643ab":"Multi-step automations that call many tools.","oss-hedl-value-ledger.e17dc4ce85":"Agent pipelines","page-breadcrumb.c766e66518":"Breadcrumb","page-toc.7e439c353e":"On this page","section_st_a1_display-split.7c9a7c0610":"Detail","section_st_f1_cta-dark.90e40d5043":"Get started","toc-rail.f5cbdf6bfb":"Contents"}