PicoClaw users report a focused chat box can spike CPU

sipeed/picoclaw issue #3292 is the inspected primary source: “[BUG] CPU usage too high when focus on input box in chat interface / 聊天界面输入框在选中时cpu占用高.” The two related records below were inspected as supporting context, not treated as independent confirmation.

The facts

- sipeed/picoclaw issue #3292 was created 2026-07-24T09:51:28Z and was open at inspection; its title is “[BUG] CPU usage too high when focus on input box in chat interface / 聊天界面输入框在选中时cpu占用高.” - sipeed/picoclaw PR #3273 was created 2026-07-20T16:15:43Z and was open at inspection; its title is “feat(webui): add Japanese (ja) localization.” - sipeed/picoclaw PR #3271 was created 2026-07-20T13:56:44Z and was open at inspection; its title is “chore(providers): update default model names to 2026-07 latest.” - The repository reported 29790 stars, 56 open issues, default branch main, and last push 2026-07-30T18:32:36Z. - The latest tagged-release baseline checked was nightly published 2026-07-02T01:26:53Z.

What changed

sipeed/picoclaw issue #3292 reports: Quick Summary CPU usage too high when focus on input box in chat interface / 聊天界面输入框在选中时cpu占用高 Environment & Tools **PicoClaw Version:** 0.3.1 **Go Version:** 1.26 **AI Model & Provider:** deepseek-v4-flash **Operating System:** debian/linux x64 **Channels:** web in firefox Steps to Reproduce 1. Run picoclaw-launcher 2. Open http://localhost:18800/ in browser 3. Make multi-cycle conversation with agent 4. You will find browser is being slow to send message to agent and cpu usage about 100% Actual Behavior The input state lives inside ChatPage component — every keystroke triggers a full re-render of the entire message list, including all react-markdown and code highlighting, causing ~100% CPU

sipeed/picoclaw PR #3273 reports: Summary Adds Japanese localization to the PicoClaw WebUI, as requested in #3272. Add web/frontend/src/i18n/locales/ja.json — full Japanese translation of en.json (all 968 lines) Register the ja resource and dayjs/locale/ja in i18n/index.ts, including the languageChanged dayjs locale switch Add a 日本語 entry to the language switcher dropdown in app-header.tsx Verification Key structure and {{placeholder}} interpolation variables programmatically verified to match en.json exactly (no missing / extra keys, no placeholder mismatches) Built the frontend and picoclaw-launcher (linux/amd64) and confirmed the UI renders in Japanese via the language switcher and browser auto-detection Closes #3272 Generated with [Claude Code]

sipeed/picoclaw PR #3271 reports: Summary Refresh default/common model lists across 9 providers to reflect current model IDs, each verified against official provider documentation **OpenAI**: gpt-5.4/gpt-5.5 → gpt-5.6-terra/gpt-5.6-luna/gpt-5.6-sol (verified via platform.openai.com/api/docs/models) **Anthropic**: claude-opus-4-7/claude-sonnet-4-6 → claude-fable-5/claude-opus-4-8/claude-sonnet-5 (verified via platform.claude.com/docs) **Gemini**: gemini-3-flash-preview → gemini-3.5-flash **DeepSeek**: deepseek-chat → deepseek-v4-flash (deepseek-chat deprecated 2026-07-24 per api-docs.deepseek.com) **Qwen**: qwen3.6-* → qwen3.7-max/qwen3.7-plus, qwen3.6-flash **Moonshot**: kimi-k2.5/kimi-k2-thinking → kimi-k3/kimi-k2.7-code/kimi-k2.6 (verified via platform.kimi.com/docs) **Zhipu**: glm-5/glm-4-flash → glm-5.2/glm-4.7-flash (verified via docs.bigmodel.cn) **MiniMax**: MiniMax-M2.5 → MiniMax-M3 **LongCat**: LongCat-Flash-Thinking → LongCat-2.0 Changes Updates Web UI provider catalog (CommonModels()), DefaultConfig() presets, auth/OAuth hardcoded defaults, legacy migration maps, config.example.json, security examples, skill examples, and all language docs (en/zh/ja/ko/pt-br/vi/fr/it/id/ms). Pure string replacement — no logic or

Why it matters

An idle input field should not become a compute workload. Focus-driven render loops waste battery and server capacity, and localization or model-catalog growth can hide the regression behind more state updates. This is an operator boundary: persisted state, execution ownership and visible controls must describe the same event. A workaround that broadens access, deletes state or hides the symptom would trade a visible defect for an unmeasured one.

Current

The primary record was open when captured on 2026-08-01. The release baseline was nightly published 2026-07-02T01:26:53Z. Current repository metadata, pull requests, issues, release records, Google News discovery, Hacker News discovery, Lobsters, Metamesh and the rendered ClawCharts row were checked. ClawCharts selected the subject; it did not prove the claim.

Evidence

The primary URL is https://github.com/sipeed/picoclaw/issues/3292. Supporting records are https://github.com/sipeed/picoclaw/pull/3273 and https://github.com/sipeed/picoclaw/pull/3271. Their source bodies, timestamps and states were captured in the daily evidence bundle. Test counts, reproductions and deployment observations remain attributed to their authors unless explicitly identified as independently rerun.

Source boundary

Open work is described as open, closed work as closed, and operator reports as reports. A pull request is evidence of proposed or reviewed direction, not proof of a shipped release. Issue closure would not by itself prove a deployed fix. Search residue, package mirrors and historically published source spines were excluded.

Operator take

Profile focus and typing separately, inspect timers and reactive subscriptions, test the Japanese UI, and verify model-name refreshes do not trigger repeated reconciliation. Preserve a before-state receipt, make the smallest reversible change, and verify the original failure independently after intervention.

Caveat

Public project records are mutable. Status, scope and evidence can change after publication. This brief records the inspected state and does not authorize changes to a reader’s deployment.

Source inspected; source state, environment and release boundary remain explicit.