- openclaw 2026.7.1 — 2026.7.1 OpenClaw v2026.7.1 brings major Control UI and onboarding overhauls, major updates to the official iOS, Android, and macOS apps, expanded model and provider support including GPT-5.6...
- fix(clickclack): add timeouts to REST client requests — What Problem This Solves Fixes an issue where users relying on the ClickClack plugin could have sends, message fetches, or activity REST calls wait indefinitely when the ClickClack API accepted the...
- refactor(media): use qrcode buffer renderer — What Problem This Solves QR PNG rendering asked qrcode for a data URL, checked and sliced the prefix, then decoded that base64 again when writing a temporary file. OpenClaw also duplicated input...
- fix(signal): route username targets as usernames, not phone numbers — Summary resolveSignalOutboundTarget (extensions/signal/src/outbound-session.ts) handled the documented username:/u: Signal send grammar by stripping the prefix to the bare name — and then fell...
- fix(feishu): explain content-policy delivery rejections — Closes #105976 What Problem This Solves Fixes an issue where Feishu users would receive the generic “reply completed without visible content” fallback when Feishu rejected the assistant reply with...
- refactor(skills): trim internal exports — What Problem This Solves The production Knip ratchet still carried 18 internal src/skills exports with no production consumers. Why This Change Was Made - Routes discovery, filtering, watcher,...
- perf: avoid event-loop stall during embedded_run bootstrap-context — Summary Fix event-loop stalls during embedded_run bootstrap-context that block the event loop for 14-22 seconds, causing message loss. Root causes addressed 1. Synchronous file I/O —...
- refactor(terminal): unify bounded buffers — Related: #107214 What Problem This Solves The embedded-terminal stack had several separate bounded queues for data waiting on a consumer. Their duplicated capacity and overflow machinery made...
- feat(setup): rename Crestodian to OpenClaw system agent — What Problem This Solves Phase 0 of #107237. "Crestodian" doesn't communicate what it is — users don't get that it's the system itself they're talking to. Before building agent→system-agent...
- fix(build): add @openclaw/ai workspace package to tsdown build config — What Problem This Solves packages/ai ships without a scripts.build field and was not added to the tsdown.config.ts configs array. After pnpm build completes successfully, packages/ai/dist/ is never...
- fix(anthropic): accept "cli" entrypoint in Claude session catalog discovery — What Problem This Solves Claude Code v2.x writes entrypoint: "cli" in its JSONL session files. OpenClaw's discoverCliRecords function in extensions/anthropic/session-catalog.ts only recognized...
- refactor(whatsapp): let SDK own inbound command context — What Problem This Solves WhatsApp currently reconstructs command turns, command sources, inbound-context finalization, source-reply inputs, and text-debounce eligibility that are already canonical...
- fix(agents): repair interrupted tool calls at provider-bound replay seam — What Problem This Solves Interrupted sessions (process crash, /new, or network drop during tool execution) leave orphaned tool_use blocks in persisted session history, causing Anthropic API to reject...
- fix: preserve Codex quota when status uses default runtime — Closes #105184 AI-assisted: Yes (Codex) What Problem This Solves Fixes an issue where Codex app-server users would lose the usage/quota line from /status when the session still used the codex...
- fix(config): MCP disabled: true is silently ignored when spawning servers — Closes #103954 What Problem This Solves Fixes an issue where users who set "disabled": true on an MCP server entry in openclaw.json would see the gateway spawn the server process anyway. The...
- fix(elevenlabs): validate baseUrl with user-friendly error for malformed endpoints — Closes #106943 Summary Problem: 用户设置格式错误的 ElevenLabs baseUrl 时，会在调用链深处收到 new URL() 抛出的原始 TypeError: Invalid URL，无法确定是 ElevenLabs baseUrl 配置的问题。 Solution: 在 normalizeElevenLabsBaseUrl 中添加 URL 格式校验，遵循...
- [Performance] Vitest OOMs from Vitest memory leaks during parallel E2E agent suites — Description Running pnpm test:extensions locally or in CI heavily leaks memory, eventually resulting in an Out Of Memory (OOM) crash. Profiling indicates that the Vitest runner caches the entire...
- fix(ui): restore normal content scrollbars and prevent chrome selection — Closes #107313 What Problem This Solves Fixes an issue where Control UI users had an undersized scrollbar across the entire app instead of only in compact sidebars. Dragging across navigation,...
- fix(codex): guard sandbox http/request JSON.parse against non-JSON stdout — What Problem This Solves runSandboxHttpRequest in extensions/codex/src/app-server/sandbox-exec-server/http.ts parsed the sandbox shell helper's stdout with an unguarded JSON.parse. When the helper...
- [Feature]: Reuse release evidence for changelog-only candidates — Summary Add trusted release evidence reuse for audited candidate deltas. Problem to solve Release candidates currently restart broad release matrices after small follow-up changes even when most...
- fix(release): stop invalid candidates before validation fan-out — What Problem This Solves Resolves a release workflow problem where Full Release Validation could dispatch the full expensive matrix and complete successfully, only for npm preflight to reject the...
- fix(release): bind npm publishing to approved package bytes — Related: #103423 Supports: #103898 What Problem This Solves Fixes an issue where npm release validation could prove one prepared package candidate while publication and Telegram prepublish evidence...
- fix(release): reuse evidence for changelog-only candidates — Closes #103423 Depends on #103809 for the bounded immutable Actions artifact reader. This PR fails closed until that reader lands. Supersedes #103632 with a focused evidence-only implementation. What...
- fix(release): reconcile contribution provenance — Closes #103429 > Draft/HOLD: do not merge until the immutable historical reconciliation passes > on exact signed head 94ff88960d7b52409c4dc9ce75e274fa73f4deac. What Problem This Solves Fixes an issue...
- [codex] Keep ambiguous delivery attempts recoverable — Summary - Keeps outbound delivery entries in ambiguous post-send states pending when adapter reconciliation is unavailable. - Preserves the existing no-blind-replay behavior while leaving the entry...
- feat(mcp): requester-scoped MCP server connections — What Problem This Solves Closes #106229. In multi-user channels (Slack workspaces, group chats), every sender shares one MCP connection set: MCP servers come only from static config/bundle manifests...
- fix(android): stop killing long chat runs with hardcoded timeouts (#106863) — What Problem This Solves Any agent turn longer than ~2 minutes started from the Android app ends in an error, while the identical turn over Telegram completes fine. Full analysis in #106863. Two...
- Renderer placeholder leak: `(see attached image)`, `Image reply`, `Audio reply`, `Media reply` leak to human-visible Telegram reply under specific content-classification paths — Summary When the chat-reply-media pipeline classifies a response as "has media blocks but no visible text" AND there's a reply directive prefix, the synthetic text placeholder used for the LLM prompt...
- OpenClaw repo stats for 2026-07-14 — 382890 stars, 80370 forks, 2887 contributors
- OpenClaw sponsors (106 total) — 106 sponsors supporting OpenClaw
- perf: bound publisher profile reads (#3073)
- perf: batch home publisher hydration (#3070)
- fix(web): defer detail histories until needed (#3072)
- fix(web): defer offscreen home images (#3071)
- perf: load plugin catalog during SSR (#3062)
- fix(api): decouple rate-limit metadata writes (#3063)
- fix: polish shared controls and toasts
- openclaw v2026.7.1 on npm — Multi-channel AI gateway with extensible messaging integrations
- openclaw/openclaw on Docker Hub — Official openclaw/openclaw docker images.
- mradermacher/Medina-Qwen3.5-27B-OpenClaw-Uncensored-i1-GGUF
- OpenClaw on DigitalOcean — Deploy your own personal AI assistant with OpenClaw on a DigitalOcean Droplet® server, a powerful open-source platform that runs entirely on your infrastructure. OpenClaw connects to the messaging...
- Google restricting Google AI Pro/Ultra subscribers for using OpenClaw
- OpenClaw is changing my life
- Show HN: Klaus – OpenClaw on a VM, batteries included
- Business Automation Architect: Turn Your AI Agent Into an Automation Engine — Most automation advice assumes you're willing to pay for Zapier or learn n8n. This skill takes a different angle: your AI agent is already capable of running workflows on its own.
- Home AI Server: Build a Private, Always-On Assistant at Home
- OpenClaw Just Shipped Native Session Scroll-Back and a Settings-Search Hardening Sweep — 10 commits landed across OpenClaw in the last 3 hours. The headline is a real chat feature — native...
- OpenClaw Just Shipped UI Polish, an Animated Claw Banner, and a WebRTC Timeout Fix — 10 Commits Land — 10 commits landed across OpenClaw in the last 3 hours. This is a UI-and-feel batch: Peter Steinberger...
- OpenClaw Just Shipped a Batch of Messaging and Request-Policy Fixes — 10 Commits Land — 10 commits landed across OpenClaw in the last 3 hours. This is a correctness-heavy batch — no flashy...
- OpenClaw Just Shipped a Wave of Refactors and ClawHub-Aware Updates — 10 Commits Land — 10 commits landed across OpenClaw in the last 3 hours. The batch is dominated by refactors that tidy...
- Meet Lemuel Gulliver: IdentyClaw Concierge on every channel — verify her HOLA before you trust — Portable identity for AI agents should feel as natural as talking to a helpful concierge — and as...
- OpenClaw Adds a Claude Session Fleet and Production Cloud Workers — OpenClaw took a clear step this window from "agent you run locally" toward "agent you run as...
- Run OpenClaw: Turn Your Raspberry Pi Into an Agentic AI Sidekick — Run OpenClaw Turn Raspberry: OpenClaw lets an AI agent act on its own, and a Raspberry Pi is the safest sandbox to run it. Here is what the build needs and why the GPIO pins matter.
- OpenClaw and ContentStudio: How I Ran My Whole Social Media Workflow From One Chat Message — AI agents are moving from answering questions to actually doing the work. Here is what that shift looks like when you put it to the test.Continue reading on Medium »
- Build Your Own Self-Hosted AI Copilot from Scratch the old way: The Complete OpenClaw Setup — A genuine ground-up build. You have a fresh Linux VM and an SSH prompt; by the end you have a private AI copilot that reaches you on…Continue reading on Medium »
- From OpenClaw to Codex — Continue reading on Medium »
- Maid (AI) — Maid is a cross-platform Flutter app for interfacing with GGUF / llama.cpp models locally, and with Ollama and OpenAI models remotely.
- ChatSonic — Combining Google's knowledge graph, ChatSonic delivers real-time data, voice interaction, AI-produced visual content, personalized avatars, and API for services. It revisits past dialogs for informed...
- Helpful Skills or Hidden Payloads? Bitdefender Labs Dives Deep into the OpenClaw Malicious Skill Trap — With hundreds of malicious OpenClaw skills blending in among legitimate ones, manually reviewing every script or command isn’t realistic — especially when skills are designed to look helpful and...
- @garrytan @ctjlewis Yep. I have an enterprise deployment with 500+ gateways one for each employee and it has become the
- RT @gregisenberg: Grok 4.5 might be the BEST model to run inside Hermes or OpenClaw RIGHT NOW. I've been sleeping on Gr
- 24% những người cho rằng tìm kiếm khách hàng là rào cản lớn nhất sẽ nhận được nhiều hơn chỉ một chương trong tài liệu.
- OpenClaw v.0.1 beta by theheroGAC has been updated on Brewology! Check out the latest version and download here: https
- 고객 확보를 가장 큰 과제로 꼽은 24%를 위해, 단순한 한 챕터 이상의 내용을 준비했습니다. GOAT CMO @0x1164가 OpenClaw Builder Bootcamp에서 「From Demo to Demand
- https://t.co/ycj2wtUuHI
- RT @david0520782123: 从 OpenClaw 直写数据库，看 AI 时代企业架构的真正缺口 最近关于 OpenClaw 这类 Agent 直接操作企业系统的讨论，很多集中在一个问题： 一个 AI Agent 是否应该拥
- @zarruk_ @openclaw Exactly..
- RT @openclaw: OpenClaw v2026.7.1 is out, with 3,063 contributions from 532 contributors! • Web UI and onboarding overha
- @asaio87 @jackfriks i mean hype cycles like openclaw can give you early momentum tho the thing that worked for postiz w
- RT @chrismatthieu: My new @RealSenseai + #ROS robot is ALIVE! Now I need to see if I can get @AgenticROS + #OpenClaw ru
- @MiaAI_lab @NVIDIAAI Or 2 GX10 then you run GLM 5.2 via colibri, in my case 1 GX10 run GLM 4.5 air fine which handle all
- If you're new to the @AskVenice ecosystem, here's the simple breakdown: Venice is a privacy-first, uncensored AI platfo
- @openclaw I'm at 4.21 and an update will break my system because its heavily modded. Suggestion for the name of my fork?
- @donedall @openclaw You were right. Hard crash on update. Main configuration recovery/reversion did not solve the crash.
- RT @openclaw: OpenClaw v2026.7.1 is out, with 3,063 contributions from 532 contributors! • Web UI and onboarding overha
- Updated my OpenClaw lobster from 6.11 to 7.1 this morning and my install got broken because of plugin issues. Claude C
- RT @GhostShivHumans: Sardine World assessment passed, Available for openclaw atlas and sardine world assessment https:
- Sardine World assessment passed, Available for openclaw atlas and sardine world assessment https://t.co/eyRnDXsc7q — Sardine World assessment passed, Available for openclaw atlas and sardine world assessment https://t.co/eyRnDXsc7q
- RT @simonw: It's been about six months since OpenClaw burst onto the scene - are you still using yours? Did it become a