AI tool comparison
Ferretlog vs Structured Output Benchmark
Which one should you ship with? Here is the side-by-side panel verdict, pricing read, reviewer split, and community vote comparison.
Developer Tools
Ferretlog
git log for your Claude Code agent runs — local, zero dependencies
50%
Panel ship
—
Community
Free
Entry
Ferretlog is a zero-dependency pure Python CLI that treats your Claude Code session logs like a git repository. It parses the raw JSONL logs in `~/.claude/projects/` and gives you git-style history browsing, diff between runs, per-tool-call breakdowns, and cost/token stats — entirely locally, with no network calls and no configuration required. If you've been using Claude Code heavily, you've likely experienced the frustration of losing track of what changed across sessions, what tools were called how many times, and how much each session actually cost across sub-agent calls. Ferretlog makes that history explorable and comparable the same way `git log` makes code history explorable. This is an indie solo project from Eitan Lebras, submitted as a Show HN. It's genuinely useful as a power-user tool for anyone doing serious Claude Code work, especially those managing multi-session agent pipelines where debugging "what did the agent do last time?" is a real pain. The zero-dependency, local-only design means there's no trust surface and no setup friction.
Developer Tools
Structured Output Benchmark
The benchmark that tests whether LLMs get JSON values right, not just syntax
75%
Panel ship
—
Community
Free
Entry
Interfaze's Structured Output Benchmark (SOB) exposes a gap that has been quietly breaking production AI pipelines: models can produce syntactically valid JSON while getting the actual values wrong. SOB measures value accuracy across 21 models using 5,000 text passages, 209 OCR documents, and 115 meeting transcripts — scoring each on seven metrics including value accuracy, faithfulness (grounding vs. hallucination), type safety, and perfect-response rate. The benchmark reveals some sobering findings. Even top models like GPT-5.4 and Claude Sonnet 4.6 achieve ~83% on text but drop to 67% on images and only 23.7% on audio. No single model dominates all modalities — GPT-5.4, GLM-4.7, Qwen3.5-35B, and Gemini 2.5 Flash cluster within one point of each other on text. Perfect response rates (all seven metrics correct) rarely exceed 50% for even the best performers. For developers building data extraction pipelines, agents that read invoices, or any system where "correct JSON" means more than syntactically valid JSON, this is required reading. The dataset is on Hugging Face, the paper is on arXiv, and the playground lets you test your own model's structured output capability directly.
Reviewer scorecard
“If you run Claude Code daily, you need this immediately. Being able to diff two sessions like git commits and see exactly which tools fired and what they cost is something that should have existed from day one. Zero-dependency Python means it just works.”
“This is the benchmark I've been waiting for. 'Valid JSON' is table stakes — the real question is whether field values are correct. This plugs a genuine gap in how we evaluate extraction pipelines.”
“This is a niche tool for a niche user (heavy Claude Code power users) and the session log format Anthropic uses is undocumented and could change at any update. Tying workflows to internal log parsing is fragile infrastructure — treat it as a convenience, not a dependency.”
“The 23.7% audio accuracy stat sounds alarming but the test data is text-normalized before scoring, meaning ASR errors are excluded. It's a better benchmark than most but the methodology choices deserve more scrutiny before you rely on it for vendor selection.”
“Agent observability tooling built by the community, not the vendor, is how this ecosystem will mature. Ferretlog is primitive but it points at a real gap: we need git-style versioning and auditability for agent sessions, not just for code.”
“No universal winner across modalities is the real story here. As agentic systems increasingly handle mixed-media inputs, this exposes that model selection needs to be task-specific. Benchmarks like SOB are how the industry gets smarter about that.”
“Terminal-only, Claude Code-specific, no visuals — this tool exists entirely outside my workflow. The underlying insight (session replay and cost tracking) is useful, but it needs a UI before it reaches anyone outside the developer community.”
“For anyone automating content workflows that extract structured data from documents, briefs, or meeting recordings, this tells you which model to actually trust for each media type. Genuinely useful before you commit to an architecture.”
Weekly AI Tool Verdicts
Get the next comparison in your inbox
New AI tools ship daily. We compare them before you waste an afternoon.