{"slug": "show-hn-loopvera-evidence-ledger-for-pipeline-debugging-no-code-yet", "title": "Show HN: LoopVera – evidence ledger for pipeline debugging (no code yet)", "summary": "LoopVera, an open-source runtime evidence ledger for debugging vision pipelines and agent workflows, is being developed to provide a see-tweak-diff-sign-off loop without replacing existing operators. The project is in early stages with no runnable code yet, aiming to bring industrial-grade debugging to open-source vision stacks and eventually support AI agent workflows.", "body_md": "**English** · [简体中文](/ProofRun/LoopVera/blob/main/README.zh-CN.md)\n\nSee inside your vision pipeline.Bring industrial-grade runtime debugging to the open-source vision stack — while exploring theruntime evidence ledger for agent workflowsunderneath.\n\nThe loop:`see every step → tweak one param & rerun → diff this run against the last → human sign-off when stakes are high`\n\n. Library-agnostic · replayable · drops into your existing code, replaces no operator.\n\n⚠️ Process open source · not a finished product · and this repo is just getting started.LoopVera is beingopen-sourced step by step, in public— first the reasoning and design, then the specs, then the code skeleton, then working slices. See the[.]RoadmapPhase 1 (Problem & Vision) is done; Phase 2 (Architecture & ADRs) is next.This repo has governance and Phase 1 narrative docs —no runnable code here yet. Starring means\"watch this direction,\"not\"production-ready.\"\n\nTest accuracy drops from 94 to 87. You sprinkle five `cv2.imwrite`\n\ncalls, your folder fills with `debug_003_v2_final.png`\n\n, and to compare two versions you just eyeball two windows — a week later you can't recall what you changed or why it \"looked fixed.\"\n\n**The algorithms aren't weak; the dev tooling is.** LoopVera targets the **see → tweak & rerun → diff across runs → human sign-off when stakes are high** layer — it does not replace your operators.\n\nToday's agents can edit code, tweak params, and batch runs — but often lack a **closed-loop substrate** underneath: intermediates scatter across chat and folders, runs don't reconcile, and high-stakes changes have no clear record of **who authorized what next**. LoopVera aims to put humans and agents on the **same evidence ledger**; the vision workbench is the first load-bearing skin, not the whole story.\n\n| Is | Is not |\n|---|---|\nA capability proof: can a workflow-state IDE + evidence/authority ledger actually be engineered? |\nAnother graph-orchestration framework or LangGraph replacement |\nVision pipeline debugging as the first load-bearing skin (direction anchor) |\nA shipped commercial vision IDE or a stable `pip` package |\nA layered experiment: orchestration swappable, ledger not (Rust director + Core verbs + `.loopvera/` ) |\nA project where \"docs complete = implementation complete\" |\n\nThis is the interaction path we are\n\nbuilding toward.`pip install loopvera`\n\nand`loopvera open`\n\narenot published yet.\n\n``` python\nimport loopvera as lv\n\ndef detect(img):\n    gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)\n    lv.observe(\"gray\", gray)                                   # one line captures the intermediate\n    _, mask = cv2.threshold(gray, lv.param(\"thresh\", 130), 255, cv2.THRESH_BINARY)\n    lv.observe(\"mask\", mask)                                   # lv.param makes this tweakable in the UI\n    return mask\npip install loopvera              # ⏳ not published — will build from source first\npython my_pipeline.py             # run once; a run is recorded automatically\nloopvera open                     # ⏳ not delivered — local browser workbench\n```\n\nRead the full pain-point + vision write-up: ** LoopVera Vision** ·\n\n[简体中文](/ProofRun/LoopVera/blob/main/docs/zh-CN/vision.md).\n\nThe\n\nLoopVera column is target design— not a claim that everything works today.\n\n`imwrite` + folders |\nad-hoc matplotlib | wire up Rerun yourself | closed commercial vision IDE | LoopVera (target) |\n|\n|---|---|---|---|---|---|\n| every step visible | ✅ | ✅ | ✅ | ||\n| tweak param & rerun | ❌ | ❌ | ❌ | ✅ | ✅ |\nside-by-side diff across runs + param diff |\n❌ | ❌ | ✅ | ||\n| library-agnostic (no operator lock-in) | ✅ | ✅ | ✅ | ❌ own only | ✅ |\nadds the loop for AI (evidence closed loop) |\n❌ | ❌ | ❌ | ✅ | |\n| price | free | free | free | 💰💰💰 | free / open source |\n\nThe differentiator isn't *\"can it show an image\"* — it's **tweak-and-rerun + cross-run diff**, without locking you into anyone's operator library.\n\nOptional. Explains what the vision workbench stands on.\n\nLoopVera stands on a **domain-neutral runtime-evidence substrate** — the same loop applies to any *\"behavior is invisible\"* domain (embedded, signal, medical, robotics):\n\n```\n① see            → ② tweak & rerun → ③ diff across runs → ④ high-stakes only:\n   capture any                                                human sign-off\n   intermediate                                                     │\n        ↑______________________________ next round ________________│\n```\n\nAdding a domain takes three hooks (`observe`\n\n/ `decode`\n\n/ `placement`\n\n); the loop itself is provided by the runtime. Vision is the first **load-bearing skin**, not the whole story — the substrate is **designed** for multiple domains (e.g. `embedded`\n\nalongside `vision`\n\n) so the loop isn't vision-only on paper; additional domains ship in later roadmap phases.\n\nDepth → [AI agent workflows](/ProofRun/LoopVera/blob/main/docs/en/agent-workflows.md) · [Why not LangGraph / all-in-one](/ProofRun/LoopVera/blob/main/docs/en/why-not-langgraph.md).\n\nLoopVera is released in stages so that **every step carries standalone value** and invites a different kind of contributor. This is deliberate: the design should be critiqued *before* the code locks it in.\n\n| Phase | What ships | Who it invites | Status |\n|---|---|---|---|\n0 · Front Door |\nREADME, governance, roadmap, contribution channels | Everyone deciding whether to watch | ✅ done |\n1 · Problem & Vision |\nPain-point narrative, vision, comparison | People who feel the same pain | ✅ done |\n2 · Architecture & ADRs |\nLayering, four pillars, decision records | Engineers who want to shape the design | ⏳ next |\n3 · Contracts (L0) |\nJSON Schema, verb catalog, conformance tests | Implementers, spec folks | ⏳ planned |\n4 · Code Skeleton |\nBuildable Rust spine, walking skeleton | Contributors who write real code | ⏳ planned |\n5 · First Vertical Slice |\n`observe → diff` demo for one library |\nEarly users, integration authors | ⏳ planned |\n6+ · Filling in |\nGate/authority, more integrations, Studio UI, pip | The whole community | ⏳ planned |\n\nFull detail, entry points, and definitions of done: ** ROADMAP.md**.\n\n— for[LoopVera Vision](/ProofRun/LoopVera/blob/main/docs/en/vision.md)**vision / CV engineers**: problem, loop, comparison, honest status.— for[AI agent workflows](/ProofRun/LoopVera/blob/main/docs/en/agent-workflows.md)**AI agent developers**: the evidence ledger your agents are missing.— for[Pipeline debugging pain points](/ProofRun/LoopVera/blob/main/docs/en/pipeline-debugging-pain-points.md)**anyone tuning pipelines**: a problems-only discussion piece.— for[Why not LangGraph / all-in-one](/ProofRun/LoopVera/blob/main/docs/en/why-not-langgraph.md)**agent-infra**: graph orchestration vs. an evidence/authority ledger.\n\nAll docs are bilingual — see the [docs index](/ProofRun/LoopVera/blob/main/docs/README.md).\n\nThis is **build-in-public** capability exploration, not a commercial launch. There is **no fixed ship date**.\n\n- ⭐\n— follow the direction as it evolves.[Star / Watch on GitHub](https://github.com/Proofrun/LoopVera) - 💬\n— real pipeline pains and architecture RFCs. This is the[GitHub Discussions](https://github.com/Proofrun/LoopVera/discussions)**normative source of truth** for decisions. - 🎮\n— chat & build help (not normative; conclusions go to Discussions).[Discord · Open Build](https://discord.gg/7Xfz8fCmE) - 📖\n— see labels like[Issues](https://github.com/Proofrun/LoopVera/issues)`good-first-issue`\n\n,`help-wanted`\n\n, and`design-rfc`\n\n. - 🤝\n**Read** before opening a PR, and the[CONTRIBUTING.md](/ProofRun/LoopVera/blob/main/CONTRIBUTING.md)[Code of Conduct](/ProofRun/LoopVera/blob/main/CODE_OF_CONDUCT.md).\n\nEarly contributors' feedback **directly shapes** which happy path we wire first.\n\n[MIT](/ProofRun/LoopVera/blob/main/LICENSE) © 2026 LoopVera / Proofrun.\n\n*See every step. Diff every run. Orchestration swappable. Workers swappable. Ledger not.*", "url": "https://wpnews.pro/news/show-hn-loopvera-evidence-ledger-for-pipeline-debugging-no-code-yet", "canonical_source": "https://github.com/Proofrun/LoopVera", "published_at": "2026-07-10 15:23:51+00:00", "updated_at": "2026-07-10 15:35:25.925484+00:00", "lang": "en", "topics": ["developer-tools", "computer-vision", "ai-agents"], "entities": ["LoopVera"], "alternates": {"html": "https://wpnews.pro/news/show-hn-loopvera-evidence-ledger-for-pipeline-debugging-no-code-yet", "markdown": "https://wpnews.pro/news/show-hn-loopvera-evidence-ledger-for-pipeline-debugging-no-code-yet.md", "text": "https://wpnews.pro/news/show-hn-loopvera-evidence-ledger-for-pipeline-debugging-no-code-yet.txt", "jsonld": "https://wpnews.pro/news/show-hn-loopvera-evidence-ledger-for-pipeline-debugging-no-code-yet.jsonld"}}