{"slug": "axon", "name": "Axon", "category": "ai", "category_name": "AI models and assistants", "description": "", "recommenders": 3, "commenters": 5, "warners": 1, "score": 1.884, "rising": -0.98, "first": "2026-02-12", "latest": "2026-08-04", "why": "I love the \u201cmandatory approval + audit log\u201d stance", "why_is_quote": true, "has_verdict": false, "sentiment": {"recommend": 3, "neutral": 1, "compare": 0, "warn": 1}, "spark": [0, 0, 0, 2, 1, 0, 0, 0, 0, 0, 0, 0], "listed": true, "loved": [], "caveat": "", "short_quote": "I love the \u201cmandatory approval + audit log\u201d stance", "kind": "", "icon": true, "domain": null, "github_repo": "github.com/axon-core/axon", "official_url": "https://github.com/axon-core/axon", "pricing": "", "aliases": ["Axon", "AXON", "axon"], "verdict": null, "praise": [{"label": "", "summary": "", "quotes": [{"id": 761357424, "item_id": 47584839, "url": "https://news.ycombinator.com/item?id=47584839", "by": "colwont", "date": "2026-03-31", "stance": "recommend", "quote": "This reduces token usage because it asks the model to think in AXON", "excerpt": "This reduces token usage because it asks the model to think in AXON https://colwill.github.io/axon", "context": "", "use_case": "Reducing output token usage", "switched_from": "", "story": "Universal Claude.md \u2013 cut Claude output tokens by 63%"}, {"id": 753846656, "item_id": 47115416, "url": "https://news.ycombinator.com/item?id=47115416", "by": "YaraDori", "date": "2026-02-22", "stance": "recommend", "quote": "I love the \u201cmandatory approval + audit log\u201d stance", "excerpt": "I love the \u201cmandatory approval + audit log\u201d stance \u2014 feels like the only sane way to run agents near real accounts. Curious how you think about the reliability layer for UI work: - do you store any per-step evidence (DOM snapshot/screenshot) in the audit trail? - do you support \u201ccheckpoints\u201d so an operator can approve once per milestone vs every click? - any redaction strategy for screenshots/logs (PII, tokens) befor", "context": "", "use_case": "agent approval and audit logging", "switched_from": "", "story": "Show HN: Axon \u2013 Agentic AI with mandatory user approval and audit logging"}, {"id": 752388672, "item_id": 47024292, "url": "https://news.ycombinator.com/item?id=47024292", "by": "gjkim", "date": "2026-02-15", "stance": "recommend", "quote": "fix the issue and open a PR", "excerpt": "axon run -p \u201cfix the issue and open a PR\u201d Self-development settings: https://github.com/axon-core/axon/blob/main/self-development/README.md", "context": "", "use_case": "", "switched_from": "", "story": "Safely run Claude \u2013\u2013dangerously-skip-permissions on Kubernetes"}]}], "caveats": [{"label": "", "summary": "", "quotes": [{"id": 786765041, "item_id": 49172815, "url": "https://news.ycombinator.com/item?id=49172815", "by": "metamet", "date": "2026-08-04", "stance": "warn", "quote": "resistance to Flock and Axon", "excerpt": "Seems these days the only cross-partisan agreement is resistance to Flock and Axon. People don't want Big Brother. Politicians unironically trying to positively refer to them as an enhanced Big Brother is completely tone deaf. And we know that police departments were working with ICE during Metro Surge: > But the Star Tribune found law enforcement agencies around the country searched several Twin Cities suburbs\u2019 came", "context": "", "use_case": "", "switched_from": "", "story": "All of Winona Police Department's Flock cameras cut down and stolen"}]}], "who": [], "tags": [], "use_cases": ["reducing output token usage", "agent approval and audit logging"], "compared": [], "makers": [{"id": 779833184, "item_id": 48739574, "url": "https://news.ycombinator.com/item?id=48739574", "by": "celltalk", "date": "2026-06-30", "stance": "recommend", "quote": "I basically did the same thing almost one and half years ago", "excerpt": "I basically did the same thing almost one and half years ago and not many people cared, but I still believe that this is the future for computational biology. https://celvox.co/solutions/axon", "context": "", "use_case": "Computational biology tool", "switched_from": "", "story": "Claude Science"}, {"id": 754440560, "item_id": 47152535, "url": "https://news.ycombinator.com/item?id=47152535", "by": "gjkim042", "date": "2026-02-25", "stance": "neutral", "quote": "todo issue to reference kubernetes secret from mcp config", "excerpt": "About handling secret/credential scoping, there\u2019s todo issue to reference kubernetes secret from mcp config. https://github.com/axon-core/axon/issues/313", "context": "", "use_case": "", "switched_from": "", "story": "Show HN: Axon \u2013 A Kubernetes-native framework for AI coding agents"}, {"id": 753058000, "item_id": 47066125, "url": "https://news.ycombinator.com/item?id=47066125", "by": "gjkim042", "date": "2026-02-18", "stance": "recommend", "quote": "I built Axon because I wanted to run Claude Code fully autonomously but didn't want to give it free reign over my laptop.", "excerpt": "Hello HN, I built Axon because I wanted to run Claude Code fully autonomously but didn't want to give it free reign over my laptop. The insight was simple: Kubernetes already solves the hard problems. Isolation? Pods. Ephemeral environments? Jobs. Credential scoping? RBAC + Secrets. Scheduling dozens of agents? That's just scheduling. So --dangerously-skip-permissions inside an ephemeral Pod isn't actually dangerous ", "context": "", "use_case": "running coding agents on Kubernetes", "switched_from": "", "story": "Show HN: Axon \u2013 Run autonomous coding agents(Claude, Codex) safely on Kubernetes"}], "threads": [{"url": "https://news.ycombinator.com/item?id=49171656", "title": "All of Winona Police Department's Flock cameras cut down and stolen", "mentions": 1}, {"url": "https://news.ycombinator.com/item?id=48735770", "title": "Claude Science", "mentions": 1}, {"url": "https://news.ycombinator.com/item?id=47581701", "title": "Universal Claude.md \u2013 cut Claude output tokens by 63%", "mentions": 1}, {"url": "https://news.ycombinator.com/item?id=47137491", "title": "Show HN: Axon \u2013 A Kubernetes-native framework for AI coding agents", "mentions": 1}, {"url": "https://news.ycombinator.com/item?id=47062684", "title": "Show HN: Axon \u2013 Agentic AI with mandatory user approval and audit logging", "mentions": 1}, {"url": "https://news.ycombinator.com/item?id=47066093", "title": "Show HN: Axon \u2013 Run autonomous coding agents(Claude, Codex) safely on Kubernetes", "mentions": 1}], "sentiment_by_month": {"2025-11": {}, "2025-12": {}, "2026-01": {}, "2026-02": {"neutral": 1, "recommend": 2}, "2026-03": {"recommend": 1}, "2026-04": {}, "2026-05": {}, "2026-06": {}, "2026-07": {}, "2026-08": {"warn": 1}, "2026-09": {}, "2026-10": {}}, "quotes": [{"id": 786765041, "item_id": 49172815, "url": "https://news.ycombinator.com/item?id=49172815", "by": "metamet", "date": "2026-08-04", "stance": "warn", "quote": "resistance to Flock and Axon", "excerpt": "Seems these days the only cross-partisan agreement is resistance to Flock and Axon. People don't want Big Brother. Politicians unironically trying to positively refer to them as an enhanced Big Brother is completely tone deaf. And we know that police departments were working with ICE during Metro Surge: > But the Star Tribune found law enforcement agencies around the country searched several Twin Cities suburbs\u2019 came", "context": "", "use_case": "", "switched_from": "", "story": "All of Winona Police Department's Flock cameras cut down and stolen"}, {"id": 761357424, "item_id": 47584839, "url": "https://news.ycombinator.com/item?id=47584839", "by": "colwont", "date": "2026-03-31", "stance": "recommend", "quote": "This reduces token usage because it asks the model to think in AXON", "excerpt": "This reduces token usage because it asks the model to think in AXON https://colwill.github.io/axon", "context": "", "use_case": "Reducing output token usage", "switched_from": "", "story": "Universal Claude.md \u2013 cut Claude output tokens by 63%"}, {"id": 753846656, "item_id": 47115416, "url": "https://news.ycombinator.com/item?id=47115416", "by": "YaraDori", "date": "2026-02-22", "stance": "recommend", "quote": "I love the \u201cmandatory approval + audit log\u201d stance", "excerpt": "I love the \u201cmandatory approval + audit log\u201d stance \u2014 feels like the only sane way to run agents near real accounts. Curious how you think about the reliability layer for UI work: - do you store any per-step evidence (DOM snapshot/screenshot) in the audit trail? - do you support \u201ccheckpoints\u201d so an operator can approve once per milestone vs every click? - any redaction strategy for screenshots/logs (PII, tokens) befor", "context": "", "use_case": "agent approval and audit logging", "switched_from": "", "story": "Show HN: Axon \u2013 Agentic AI with mandatory user approval and audit logging"}, {"id": 752388672, "item_id": 47024292, "url": "https://news.ycombinator.com/item?id=47024292", "by": "gjkim", "date": "2026-02-15", "stance": "recommend", "quote": "fix the issue and open a PR", "excerpt": "axon run -p \u201cfix the issue and open a PR\u201d Self-development settings: https://github.com/axon-core/axon/blob/main/self-development/README.md", "context": "", "use_case": "", "switched_from": "", "story": "Safely run Claude \u2013\u2013dangerously-skip-permissions on Kubernetes"}, {"id": 751901024, "item_id": 46993814, "url": "https://news.ycombinator.com/item?id=46993814", "by": "chaps", "date": "2026-02-12", "stance": "neutral", "quote": "this is true for any agency that uses axon", "excerpt": "Sure -- it's something I figured out during the 2020 protests for some reporting work I was doing which led to this reporting: https://thetriibe.com/2020/12/hundreds-of-chicago-police-made-arrests-without-body-cameras-during-2020-uprisings/ This reporting was made possible because it's surprisingly easy to export recording start/stop time, file size, duration, notes, cop badge and model name from the underlying syste", "context": "", "use_case": "", "switched_from": "", "story": "Text classification with Python 3.14's ZSTD module"}]}