{"slug": "show-hn-openclacky-a-token-efficient-personal-agent-written-in-ruby", "title": "Show HN: OpenClacky – A token-efficient personal agent written in Ruby", "summary": "OpenClacky, a token-efficient open-source AI agent written in Ruby, claims to deliver Claude Code-class capability at a lower cost than competing agents. The system reports a 94.9% cache hit rate and has completed over 488,000 tasks, with developers citing up to 85% savings on input bills compared to rivals.", "body_md": "# The Most Token-Efficient Open-Source AI Agent\n\nClaude Code-class capability. Lower cost than any Agent in its class.\n\n**94.9%**·\n\n**488,424** tasks completed\n\n**5 decisions** and\n\n**8 tasks**, assigned by owner: Mia 3 (interaction draft / user-interview script / event-tracking audit), Leo 3 (performance regression / rollout checklist / release schedule), Kai 2 (customer commitments follow-up / reconciliation). Results written to\n\n`tasks/2026-W18.md`\n\n. Want me to sync to the team calendar?## Cheapest. And Best.\n\nEngineered for the best token economics — saving you the most time and money.\n\n[\nLowest Cost\nUS\n~0.8×\n16 tools · chasing 100% cache hit · smart subagent routing\n1.0×\nWorld-class harness, closed-source, subscription\n~1.5×\nPeer harness agent\n~3–4×\n52 built-in tools, schema bloat\n](/benchmark)\n\nHead-to-head with the strongest rivals, we cost less. At equal capability, OpenClacky's token spend stays the lowest.\n\nCost comparisons are computed from the same model provider, using identical or capability-equivalent models billed via API.\n\nSee Full Cost Report →[\nTop Quality\nvs Claude Code & Codex\nCompletion\nExcellent\nClarity & Layout\nEdge\nExcellent\nAI Flavor (lighter is better)\nEdge\nExcellent\nSame output, just\nWatch the Head-to-Head\n→\n](/benchmark/vs-claude-code-codex)\n\nHead-to-head with the strongest rivals, output matches and details win. Scored task-by-task from first-hand reviewer comparison.\n\nOutput is first-tier alongside Claude Code, with a slight edge in clarity, layout and motion.\n\n**1/3** the cost of Codex ($4.50 vs $12.54)\n\n## Don't take our word for it. Watch the network.\n\nNot marketing copy. These numbers come from anonymous telemetry across all OpenClacky users, refreshed hourly.\n\n**What does this mean?** About **85%** off your input bill. Cached tokens are billed at ~10% of full price (as low as 1% on DeepSeek & MiMo) — the higher the hit rate, the more you save.\n\n## Saves tokens · Works with any model · Easy to install\n\nReal feedback from real users, unedited.\n\nSame project, 3 hours of coding — 8× fewer tokens than Hermes.\n\nUnder 100 MB install vs Hermes at 48 GB. The WebUI is fantastic. After using this, I realized multi-agent is unnecessary. Our entire cross-border e-commerce team is on it now.\n\nCodex with the most expensive model couldn't solve it in 30 minutes. Clacky did it with kimi-k2.6. Windows setup is a breeze too.\n\nCompared to Hermes and OpenClaw, OpenClacky's tool calling is rock solid. Works great with domestic models.\n\nIt takes so much trial and error to arrive at simplest design. You did it. This is the brilliance, clarity and elegance of ruby brought to agentic tooling. I'm blown away.\n\nCreates awesome ping pong tutorial videos with OpenClacky\n\nAll quotes are genuine user feedback, published publicly or with permission.\n\n##\nFour things done right.\n\nTokens follow.\n\nNot by cutting features — by compounding correct choices at every layer.\n\n### Ultra-High Cache Hit Rate\n\nSession never restarts. Double-window cache markers. Insert-then-Compress reuses cache even during history compression — system prompt never mutates.\n\n**approach 100%**\n\n### Minimal Toolset, Maximal Reach\n\nJust **16** core tools. Capabilities extend via the invoke_skill meta-tool. Tool count isn't competitive advantage — task completion is.\n\n### Idle-time compression. Don't pay while you're away.\n\nStep away for a meeting or coffee — the agent quietly compresses long context in the background. By the time you're back, the history loaded into the model is already down to its essence. The tokens normally burned between turns? Gone.\n\n**30%+** of total session cost\n\n### Pick your model. Pick your route.\n\nAny OpenAI-compatible API, plug and play. Official direct, aggregated routing, compatible relays — the choice is 100% yours, and so is the floor on your cost.\n\n## Pick your model. Set your cost.\n\nBring your own key — completely free. Any OpenAI-compatible API works out of the box. Or use OpenClacky Keys for the hands-off path: direct upstream routing, 99% cache hits, official pricing.\n\nNot locked to one vendor. Main task on Claude, subagents auto-route to DeepSeek — another big chunk of tokens saved.\n\n##\nDouble-click to install.\n\nNo setup required.\n\nDownload, double-click, start using — environment, dependencies, Skills all ready. Desktop installer recommended; CLI also available for power users.\n\nDouble-click to install — full environment configured automatically\n\nOne line in your terminal — for developers who prefer the CLI\n\n## Skills are the soul of any modern agent\n\nSummon them, build them, grow them, sell them — OpenClacky treats Skills as native capability, not plug-ins.\n\n`/`\n\nSummon any Skill\n\nType / to browse, fuzzy-search, and invoke instantly. Hundreds of Skills one keystroke away — no menus, no memorizing names.\n\n### Build Skills by chatting\n\nTell the agent what you want. It drafts the SKILL.md, breaks the steps down, runs validation. No code required.\n\n### Skills self-evolve\n\nAfter each run, the AI reviews context and outcome, then updates the Skill itself. Next time you invoke it, it performs more reliably.\n\n### Open formats, ready to monetize\n\nCompatible with Skill formats from any agent ecosystem. Polished Skills can be packaged, encrypted, licensed, and priced on your terms.\n\n##\nOnce you master it,\n\npackage your workflow and sell it\n\nProfessionals are already building their own vertical AI experts on OpenClacky — encrypted distribution, license management, your own pricing.\n\n###\n10 years of legal expertise,\n\nnow running on every lawyer's desktop\n\n\"I just put my 10 years of expertise into every colleague's computer.\"\n\n###\nAI life planning assistant\n\nfor every major decision\n\n\"Wealth gets divided. Life wisdom in the system stays yours.\"\n\n###\nJunior doctors access\n\nsenior expertise instantly\n\n\"The veteran retired, but the expertise stayed.\"\n\n## Open-source code. Verifiable results.\n\nAll code is public, all data is public. One month in — every number can be checked.\n\n## Make every token earn its keep\n\nOpen, free, BYOK. Download OpenClacky and take back control of your AI budget.", "url": "https://wpnews.pro/news/show-hn-openclacky-a-token-efficient-personal-agent-written-in-ruby", "canonical_source": "https://www.openclacky.com/", "published_at": "2026-06-11 20:34:56+00:00", "updated_at": "2026-06-11 20:50:39.803449+00:00", "lang": "en", "topics": ["ai-agents", "ai-tools", "ai-products", "ai-startups", "large-language-models"], "entities": ["OpenClacky", "Claude Code", "Codex", "Mia", "Leo", "Kai"], "alternates": {"html": "https://wpnews.pro/news/show-hn-openclacky-a-token-efficient-personal-agent-written-in-ruby", "markdown": "https://wpnews.pro/news/show-hn-openclacky-a-token-efficient-personal-agent-written-in-ruby.md", "text": "https://wpnews.pro/news/show-hn-openclacky-a-token-efficient-personal-agent-written-in-ruby.txt", "jsonld": "https://wpnews.pro/news/show-hn-openclacky-a-token-efficient-personal-agent-written-in-ruby.jsonld"}}