OpenClaw Newsletter - 2026-06-20
OpenClaw Newsletter
Saturday, June 20, 2026Top Stories
- docs(providers): register Ace Data Cloud third-party provider plugin — Summary Adds @acedatacloud/openclaw-provider as an external provider entry so the package becomes discoverable through openclaw plugins install and the onboarding wizard, and ships a...
- feat(web-search): expose Brave Goggles for custom search filtering and ranking — Summary - Problem: Brave Search supports Goggles, a query-time DSL that redefines how results are ranked at the index level (not client-side post-filtering), but OpenClaw's web_search tool has no way...
- fix(memory): preserve Windows QMD command paths — What Problem This Solves Fixes #92302. Windows users can configure memory.qmd.command with an absolute QMD command path such as C:\Users\...\qmd.js, but the memory backend previously passed that...
- fix: #94426 isPortBusy probes all address families via checkPortInUse — Summary - Route isPortBusy through checkPortInUse (from src/infra/ports-inspect.ts) so it probes all four endpoints (127.0.0.1, 0.0.0.0, ::1, ::) instead of binding only the IPv6 wildcard (::), which...
- message tool fails to deliver files/images on Feishu (400 volc-dcdn / write ECONNRESET) while same Lark SDK upload succeeds standalone
- fix(ci): route Markdown-only PR jobs away from Blacksmith runners — What Problem This Solves Markdown-only PRs (e.g. CONTRIBUTING.md) still queue preflight and security-fast on Blacksmith-backed runners because the docs-only detection runs inside preflight after the...
- fix(config): add stdio to McpServerSchema transport union — Summary openclaw config validate rejects transport: "stdio" on command-bearing MCP servers, even though the runtime (resolveMcpTransportConfig) correctly resolves any server with a command field as...
- fix(nodes): guard node catalog and pairing-list scalar fields against non-string entries — Summary buildEffectiveKnownNode in src/gateway/node-catalog.ts emits the node scalar fields raw from the pairing sources, while only the array fields (caps/commands) are hardened via...
- fix: bound Codex Telegram turns fail after /codex bind on OAuth refresh — What Problem This Solves Fixes an issue where users who bind a Telegram conversation to Codex with /codex bind would hit Codex app-server turn failed: missing field accessToken on the first follow-up...
- fix(memory-core): include path in FTS indexed text so filename tokens are searchable — What problem does this PR solve? openclaw memory search can rank filename/date-token queries poorly or miss them when the filename token only lives in the derived path FTS index. The current PR was...
- [codex] feat(agent): add session self-compaction action — What Problem This Solves Fixes #6757. OpenClaw can compact context automatically or through external/manual paths, but the in-run model had no first-class way to request compaction for its own active...
- fix(infra): report macOS product version in agent runtime.os — Summary - Problem: On macOS 26 (Tahoe), OpenClaw injected the Darwin kernel release from os.release() (e.g. 25.5.0) into the agent system prompt's runtime.os field. Because the Darwin and macOS...
- fix #90769: TUI markdown renderer strips underscores inside fenced code blocks, corrupting Python dunder syntax — Summary - Fix classification: Root-cause fix for the TUI markdown rendering path. - Root cause: The OpenClaw TUI rendered chat Markdown through pi-tui and then let pi-tui wrap every rendered physical...
- fix(update): preserve local package overrides — Summary - Preserves local package overrides during package update flows, including standalone added dist files, and reports the recovery bundle when overrides are not reapplied. - Keeps...
- fix(feishu): replace Type.Any() with explicit AnyJsonValue schema for Bedrock compatibility — Summary Fix three bitable write tools that use Type.Record(Type.String(), Type.Any()) which serializes to empty {} sub-schema, rejected by AWS Bedrock strict draft 2020-12 validation. Changes - Add...
- fix(channels): avoid copying process.env in ingress queue hot path (fixes #94571) — What Problem This Solves In Kubernetes pods with enableServiceLinks enabled, process.env can contain thousands of service environment variables. The Telegram isolated polling ingress queue calls...
- fix(context-engine): read allowModelOverride from plugin config instead of hardcoding false — Summary - LCM compaction repeatedly fails with "model override denied" because resolveContextEngineCapabilities hardcodes allowModelOverride: false in the context-engine authority. - When the owning...
- fix: respect hook config model override in session-memory slug generation (#89551) — Description When llmSlug is enabled in the session-memory hook config with a model override, that model was ignored during slug generation. The slug generator always used the agent's default model....
- fix #94850: remove Lobstering from default progress draft labels — Summary The default auto progress-label pool (DEFAULT_PROGRESS_DRAFT_LABELS) included Lobstering, which could appear as the top-level status in Telegram progress drafts without any user opt-in. The...
- [codex] Use Executing for default progress label — Summary - Replace the default automatic progress label Lobstering with Executing. - Update the progress-drafts documentation so the documented default pool matches runtime behavior. - Add focused...
- fix(cron): detect recovered tool warnings without preferFinalAssistantVisibleText — Summary An isolated scheduled agentTurn that recovers from an early tool error, produces final assistant output, and ends with status=success should still be eligible for delivery dispatch. Instead,...
- fix: avoid blocking stale PreToolUse relays — Summary - default unavailable Codex PreToolUse native hook relays to an observational no-op - keep unavailable PermissionRequest handling fail-closed - update native hook relay command and CLI tests...
- lossless-claw sanitizer repeatedly drops duplicate tool_use blocks from legacy null transcript rows — Summary OpenClaw/lossless-claw repeatedly logs: sanitizeToolUseResultPairing dropped 10 duplicate assistant tool_use block(s) on the active main session during context assembly. Read-only DB/session...
- fix(codex): omit idle PreToolUse hook relays — Summary - Stops Codex native hook config from installing PreToolUse when the OpenClaw relay has no local pre-tool work to do. - Keeps PermissionRequest and other requested hooks intact when they...
- feat(xai): add grok-composer-2.5-fast model — Summary - Adds grok-composer-2.5-fast to xAI model catalog in extensions/xai/model-definitions.ts - Adds grok-composer prefix to modern model list in extensions/xai/provider-models.ts - Model is live...
- docs: add Windows pnpm fallback for Corepack EPERM — Summary This PR adds a Windows-specific fallback note for Corepack EPERM errors during OpenClaw onboarding. On some Windows environments, running corepack enable fails because Corepack cannot create...
- fix(claude-cli): route native tool use through operator approver when ask is configured — Summary - Claude native tool use (Write/Edit/Read) was hard-denied for all exec policies except security="full" + ask="off" - With ask="on-miss" or ask="always", operators were never actually asked —...
- fix: route Discord CLI sends through gateway — Direct openclaw message send --channel discord currently takes the Discord outbound adapter's direct path, which bypasses gateway-scoped message_sending/message_sent hooks. That leaves runtime...
- fix(status): reduce deep status manifest scans — Summary - Reuse the gateway startup plugin metadata snapshot for compatible direct metadata loads so deep status paths avoid repeatedly restatting bundled plugin manifests. - Carry bundled manifests...
- OpenClaw repo stats for 2026-06-20 — 379572 stars, 79468 forks, 2618 contributors
- OpenClaw sponsors (108 total) — 108 sponsors supporting OpenClaw
- fix: backfill stored plugin manifests (#2763)
- fix: harden plugin manifest backfill — Supports direct config-map plugin manifests and adds a cursor-based production backfill page action.
- feat: backfill plugin manifest summaries — Adds a production-safe backfill for latest active plugin package releases so existing plugin pages can render typed manifest capability tabs without republishing.
- feat: render plugin manifest capabilities (#2759)
- fix(web): hide pending and suspicious items from New (#2760)
- Refresh homepage experience (#2756) — style: add Mobbin-inspired home listing toolbar and grid Introduce a catalog section below the hero with Skills/Plugins toggle, sort tabs, filter affordance, and a card grid fed by public list APIs....
- fix: make catalog metadata publisher-owned (#2757) — fix: make catalog topics canonical fix: make catalog metadata publisher-owned
- fix: align official plugin browse with official publishers (#2758) — fix: bound official plugin browse queries fix: normalize official publisher plugins
- fix: allow github sync owner-scoped skill slugs
- feat: add Vercel Speed Insights
- update security section
- mradermacher/Qwen3.5-9b-Opus-Openclaw-Distilled-i1-GGUF — text-generation
- 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...
- A sane but bull case on Clawdbot / OpenClaw
- You are not supposed to install OpenClaw on your personal computer
- I Trained My OpenClaw to Dream. Here's What It Learned Overnight. — Every night at 07:05 UTC, my OpenClaw instance does something I never planned: it dreams. Not...
- I Wired OpenRouter Free Models Into My OpenClaw Fallback Chain. Here's What Actually Works. — Three weeks ago my OpenClaw agent started returning overloaded_error during peak hours. Not because...
- Running a local AI agent on an 8GB Jetson Orin Nano (2026) — A practical look at running an always-on AI agent on an 8GB Jetson Orin Nano — what works, what doesn’t, and how to skip the setup pain.
- My OpenClaw Cron Broke and Fixed Itself Before I Noticed — Last Thursday I woke up to a Telegram alert: "Cron self-repair: 1 repair made." I hadn't triggered...
- My OpenClaw Security Audit Hung Every Night — A 3-Line Bash Fix and What It Taught Me About Sandboxing Agents
- arXiv — 에이전트 생성형 LLM을 위한 집단 스킬 트리 탐색 기술, OpenClaw-Skill — 1. 한눈에 보는 핵심 요약 (Executive Summary)Continue reading on Medium »
- [OpenClaw #2] Setting Up OpenClaw and Running My First Agent — From Installation to First ConversationContinue reading on Medium »
- Como Usar o OpenClaw com o WhatsApp em uma VPS — O OpenClaw é um agente de IA open source que funciona como uma central de mensagens: ele conecta WhatsApp, Telegram, Slack e outros canais…Continue reading on Medium »
- Hermes Agent: OpenClaw alternative — Last week, I watched a video on Piyush Garg’s YouTube channel about Hermes Agent. He broke it down as a solid OpenClaw alternative and…Continue reading on Medium »
- Your One-Click Private Agent Is a Chatbot Until You Make It Dangerous — Golden Jacket Field Notes, Week 14Continue reading on AWS Tip »
- Goose OSS — An open-source platform for seamless automation in engineering tasks, providing local execution and control. Customizable with any LLM for enhanced autonomy, enabling developers to efficiently handle...
- Khoj — Open-source app offers local or cloud AI chat with major LLMs, advanced semantic search, automation, personal newsletters, document indexing, image generation, and text-to-speech. Access from web,...
- 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...
- imagine trying to find original openclaw in 2026 https://t.co/twUysCdyux
- New research out of Google DeepMind (Mar 2026): "AI Agent Traps" — attackers don't need to hack the model, they just man
- Siam Piwat builds a borderless luxury ecosystem 🇹🇭 Belmond (LVMH) + Galeries Lafayette + Insignia + MJets. Retail × pr
- RT @elonmusk: People giving OpenClaw root access to their entire life https://t.co/jqW0FwiW03
- RT @OpenAcademyAI: 🦀 OpenClaw 101 — free Duolingo-style course - 8 short lessons, theory + practice - 15 minutes and
- ada yang daily nya pake Grok dari X premium buat Hermes Agent atau OpenClaw? gimana performa sama limit nya? https://t.c
- RT @reviseio: Collaborate on docs with agents like Claude, Codex, and OpenClaw.
- Autonomous agents are disrupting: customer support (instant), marketing (24/7 content), operations (zero downtime), sale
- Recurring reminder that if product is free or subsidized, even if you're paying $200/mo, YOU are the product 💀 The AI la
- @li9292 我一直在用openclaw 和hermes，智能体的失忆问题很困扰着我。每天都是不断的让他固化记忆。 EverOS真的在做大事。 — @li9292 我一直在用openclaw 和hermes，智能体的失忆问题很困扰着我。每天都是不断的让他固化记忆。 EverOS真的在做大事。
- Attention $OPENCLAW Family! YOUR vote matters! Less than 100 votes are needed to list $OPENCLAW on the Moonshot Top 100
- hermes is at 8x openclaw
- RT @BSPK_: 저만 하더라도 에이전트 프로젝트를 고정 워크플로우 자동화는 Claude Code, 동적 워크플로우는 NemoClaw(OpenClaw/Hermes), 시스템 통합은 Palantir로 나눠서 관리 중
- @theo @mweinbach @FrameworkPuter what if the crab was a lobster (openclaw)
- @Glitch_Trades Oubli openclaw tu vas penser un temps fou à corriger ses erreurs , hermes est top 👍 Commence par 1 agent
- RT @GoatVietNam: Tuần này tại GOAT Network: • Chúng tôi công bố OpenClaw Summer Builder Bootcamp 2026 với vai trò đồng
- @magmareactive Aynen çok mantıklı bir akış daha önce bunu OpenClaw üzerinden yaptırdım ama insanlar fazla karışık ve uzu
- @petrabarus @MuhammadGhifary @pebaryan @ainunnajib @iboen Jangan dipusingin bang, belom tentu kepake jg barangnya, herme
- @aronhatsuyuki ローカルLLM（AI）を導入したいのは、AIコーディングエージェント（プログラミング）や、OpenClaw（AIエージェント、代理で行わせる作業を自然言語で指示できる）を動かしてみたいんですよ 何かを作ろうと
Trending on X
- imagine trying to find original openclaw in 2026 https://t.co/twUysCdyux
- New research out of Google DeepMind (Mar 2026): "AI Agent Traps" — attackers don't need to hack the model, they just man
- Siam Piwat builds a borderless luxury ecosystem 🇹🇭 Belmond (LVMH) + Galeries Lafayette + Insignia + MJets. Retail × pr
- RT @elonmusk: People giving OpenClaw root access to their entire life https://t.co/jqW0FwiW03
- RT @OpenAcademyAI: 🦀 OpenClaw 101 — free Duolingo-style course - 8 short lessons, theory + practice - 15 minutes and
- ada yang daily nya pake Grok dari X premium buat Hermes Agent atau OpenClaw? gimana performa sama limit nya? https://t.c
- RT @reviseio: Collaborate on docs with agents like Claude, Codex, and OpenClaw.
- Autonomous agents are disrupting: customer support (instant), marketing (24/7 content), operations (zero downtime), sale
- Recurring reminder that if product is free or subsidized, even if you're paying $200/mo, YOU are the product 💀 The AI la
- @li9292 我一直在用openclaw 和hermes，智能体的失忆问题很困扰着我。每天都是不断的让他固化记忆。 EverOS真的在做大事。 — @li9292 我一直在用openclaw 和hermes，智能体的失忆问题很困扰着我。每天都是不断的让他固化记忆。 EverOS真的在做大事。
- Attention $OPENCLAW Family! YOUR vote matters! Less than 100 votes are needed to list $OPENCLAW on the Moonshot Top 100
- hermes is at 8x openclaw
- RT @BSPK_: 저만 하더라도 에이전트 프로젝트를 고정 워크플로우 자동화는 Claude Code, 동적 워크플로우는 NemoClaw(OpenClaw/Hermes), 시스템 통합은 Palantir로 나눠서 관리 중
- @theo @mweinbach @FrameworkPuter what if the crab was a lobster (openclaw)
- @Glitch_Trades Oubli openclaw tu vas penser un temps fou à corriger ses erreurs , hermes est top 👍 Commence par 1 agent
- RT @GoatVietNam: Tuần này tại GOAT Network: • Chúng tôi công bố OpenClaw Summer Builder Bootcamp 2026 với vai trò đồng
- @magmareactive Aynen çok mantıklı bir akış daha önce bunu OpenClaw üzerinden yaptırdım ama insanlar fazla karışık ve uzu
- @petrabarus @MuhammadGhifary @pebaryan @ainunnajib @iboen Jangan dipusingin bang, belom tentu kepake jg barangnya, herme
- @aronhatsuyuki ローカルLLM（AI）を導入したいのは、AIコーディングエージェント（プログラミング）や、OpenClaw（AIエージェント、代理で行わせる作業を自然言語で指示できる）を動かしてみたいんですよ 何かを作ろうと
Releases
- OpenClaw repo stats for 2026-06-20 — 379572 stars, 79468 forks, 2618 contributors
Community
- docs(providers): register Ace Data Cloud third-party provider plugin — Summary Adds @acedatacloud/openclaw-provider as an external provider entry so the package becomes discoverable through openclaw plugins install and the onboarding wizard, and ships a...
- feat(web-search): expose Brave Goggles for custom search filtering and ranking — Summary - Problem: Brave Search supports Goggles, a query-time DSL that redefines how results are ranked at the index level (not client-side post-filtering), but OpenClaw's web_search tool has no way...
- fix(memory): preserve Windows QMD command paths — What Problem This Solves Fixes #92302. Windows users can configure memory.qmd.command with an absolute QMD command path such as C:\Users\...\qmd.js, but the memory backend previously passed that...
- fix: #94426 isPortBusy probes all address families via checkPortInUse — Summary - Route isPortBusy through checkPortInUse (from src/infra/ports-inspect.ts) so it probes all four endpoints (127.0.0.1, 0.0.0.0, ::1, ::) instead of binding only the IPv6 wildcard (::), which...
- message tool fails to deliver files/images on Feishu (400 volc-dcdn / write ECONNRESET) while same Lark SDK upload succeeds standalone
- fix(ci): route Markdown-only PR jobs away from Blacksmith runners — What Problem This Solves Markdown-only PRs (e.g. CONTRIBUTING.md) still queue preflight and security-fast on Blacksmith-backed runners because the docs-only detection runs inside preflight after the...
- fix(config): add stdio to McpServerSchema transport union — Summary openclaw config validate rejects transport: "stdio" on command-bearing MCP servers, even though the runtime (resolveMcpTransportConfig) correctly resolves any server with a command field as...
- fix(nodes): guard node catalog and pairing-list scalar fields against non-string entries — Summary buildEffectiveKnownNode in src/gateway/node-catalog.ts emits the node scalar fields raw from the pairing sources, while only the array fields (caps/commands) are hardened via...
- fix: bound Codex Telegram turns fail after /codex bind on OAuth refresh — What Problem This Solves Fixes an issue where users who bind a Telegram conversation to Codex with /codex bind would hit Codex app-server turn failed: missing field accessToken on the first follow-up...
- fix(memory-core): include path in FTS indexed text so filename tokens are searchable — What problem does this PR solve? openclaw memory search can rank filename/date-token queries poorly or miss them when the filename token only lives in the derived path FTS index. The current PR was...
- [codex] feat(agent): add session self-compaction action — What Problem This Solves Fixes #6757. OpenClaw can compact context automatically or through external/manual paths, but the in-run model had no first-class way to request compaction for its own active...
- fix(infra): report macOS product version in agent runtime.os — Summary - Problem: On macOS 26 (Tahoe), OpenClaw injected the Darwin kernel release from os.release() (e.g. 25.5.0) into the agent system prompt's runtime.os field. Because the Darwin and macOS...
- fix #90769: TUI markdown renderer strips underscores inside fenced code blocks, corrupting Python dunder syntax — Summary - Fix classification: Root-cause fix for the TUI markdown rendering path. - Root cause: The OpenClaw TUI rendered chat Markdown through pi-tui and then let pi-tui wrap every rendered physical...
- fix(update): preserve local package overrides — Summary - Preserves local package overrides during package update flows, including standalone added dist files, and reports the recovery bundle when overrides are not reapplied. - Keeps...
- fix(feishu): replace Type.Any() with explicit AnyJsonValue schema for Bedrock compatibility — Summary Fix three bitable write tools that use Type.Record(Type.String(), Type.Any()) which serializes to empty {} sub-schema, rejected by AWS Bedrock strict draft 2020-12 validation. Changes - Add...
- fix(channels): avoid copying process.env in ingress queue hot path (fixes #94571) — What Problem This Solves In Kubernetes pods with enableServiceLinks enabled, process.env can contain thousands of service environment variables. The Telegram isolated polling ingress queue calls...
- fix(context-engine): read allowModelOverride from plugin config instead of hardcoding false — Summary - LCM compaction repeatedly fails with "model override denied" because resolveContextEngineCapabilities hardcodes allowModelOverride: false in the context-engine authority. - When the owning...
- fix: respect hook config model override in session-memory slug generation (#89551) — Description When llmSlug is enabled in the session-memory hook config with a model override, that model was ignored during slug generation. The slug generator always used the agent's default model....
- fix #94850: remove Lobstering from default progress draft labels — Summary The default auto progress-label pool (DEFAULT_PROGRESS_DRAFT_LABELS) included Lobstering, which could appear as the top-level status in Telegram progress drafts without any user opt-in. The...
- [codex] Use Executing for default progress label — Summary - Replace the default automatic progress label Lobstering with Executing. - Update the progress-drafts documentation so the documented default pool matches runtime behavior. - Add focused...
- fix(cron): detect recovered tool warnings without preferFinalAssistantVisibleText — Summary An isolated scheduled agentTurn that recovers from an early tool error, produces final assistant output, and ends with status=success should still be eligible for delivery dispatch. Instead,...
- fix: avoid blocking stale PreToolUse relays — Summary - default unavailable Codex PreToolUse native hook relays to an observational no-op - keep unavailable PermissionRequest handling fail-closed - update native hook relay command and CLI tests...
- lossless-claw sanitizer repeatedly drops duplicate tool_use blocks from legacy null transcript rows — Summary OpenClaw/lossless-claw repeatedly logs: sanitizeToolUseResultPairing dropped 10 duplicate assistant tool_use block(s) on the active main session during context assembly. Read-only DB/session...
- fix(codex): omit idle PreToolUse hook relays — Summary - Stops Codex native hook config from installing PreToolUse when the OpenClaw relay has no local pre-tool work to do. - Keeps PermissionRequest and other requested hooks intact when they...
- feat(xai): add grok-composer-2.5-fast model — Summary - Adds grok-composer-2.5-fast to xAI model catalog in extensions/xai/model-definitions.ts - Adds grok-composer prefix to modern model list in extensions/xai/provider-models.ts - Model is live...
- docs: add Windows pnpm fallback for Corepack EPERM — Summary This PR adds a Windows-specific fallback note for Corepack EPERM errors during OpenClaw onboarding. On some Windows environments, running corepack enable fails because Corepack cannot create...
- fix(claude-cli): route native tool use through operator approver when ask is configured — Summary - Claude native tool use (Write/Edit/Read) was hard-denied for all exec policies except security="full" + ask="off" - With ask="on-miss" or ask="always", operators were never actually asked —...
- fix: route Discord CLI sends through gateway — Direct openclaw message send --channel discord currently takes the Discord outbound adapter's direct path, which bypasses gateway-scoped message_sending/message_sent hooks. That leaves runtime...
- fix(status): reduce deep status manifest scans — Summary - Reuse the gateway startup plugin metadata snapshot for compatible direct metadata loads so deep status paths avoid repeatedly restatting bundled plugin manifests. - Carry bundled manifests...
- OpenClaw sponsors (108 total) — 108 sponsors supporting OpenClaw
- fix: backfill stored plugin manifests (#2763)
- fix: harden plugin manifest backfill — Supports direct config-map plugin manifests and adds a cursor-based production backfill page action.
- feat: backfill plugin manifest summaries — Adds a production-safe backfill for latest active plugin package releases so existing plugin pages can render typed manifest capability tabs without republishing.
- feat: render plugin manifest capabilities (#2759)
- fix(web): hide pending and suspicious items from New (#2760)
- Refresh homepage experience (#2756) — style: add Mobbin-inspired home listing toolbar and grid Introduce a catalog section below the hero with Skills/Plugins toggle, sort tabs, filter affordance, and a card grid fed by public list APIs....
- fix: make catalog metadata publisher-owned (#2757) — fix: make catalog topics canonical fix: make catalog metadata publisher-owned
- fix: align official plugin browse with official publishers (#2758) — fix: bound official plugin browse queries fix: normalize official publisher plugins
- fix: allow github sync owner-scoped skill slugs
- feat: add Vercel Speed Insights
- update security section
News
- mradermacher/Qwen3.5-9b-Opus-Openclaw-Distilled-i1-GGUF — text-generation
- 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...
- A sane but bull case on Clawdbot / OpenClaw
- You are not supposed to install OpenClaw on your personal computer
- I Trained My OpenClaw to Dream. Here's What It Learned Overnight. — Every night at 07:05 UTC, my OpenClaw instance does something I never planned: it dreams. Not...
- I Wired OpenRouter Free Models Into My OpenClaw Fallback Chain. Here's What Actually Works. — Three weeks ago my OpenClaw agent started returning overloaded_error during peak hours. Not because...
- Running a local AI agent on an 8GB Jetson Orin Nano (2026) — A practical look at running an always-on AI agent on an 8GB Jetson Orin Nano — what works, what doesn’t, and how to skip the setup pain.
- My OpenClaw Cron Broke and Fixed Itself Before I Noticed — Last Thursday I woke up to a Telegram alert: "Cron self-repair: 1 repair made." I hadn't triggered...
- My OpenClaw Security Audit Hung Every Night — A 3-Line Bash Fix and What It Taught Me About Sandboxing Agents
- arXiv — 에이전트 생성형 LLM을 위한 집단 스킬 트리 탐색 기술, OpenClaw-Skill — 1. 한눈에 보는 핵심 요약 (Executive Summary)Continue reading on Medium »
- [OpenClaw #2] Setting Up OpenClaw and Running My First Agent — From Installation to First ConversationContinue reading on Medium »
- Como Usar o OpenClaw com o WhatsApp em uma VPS — O OpenClaw é um agente de IA open source que funciona como uma central de mensagens: ele conecta WhatsApp, Telegram, Slack e outros canais…Continue reading on Medium »
- Hermes Agent: OpenClaw alternative — Last week, I watched a video on Piyush Garg’s YouTube channel about Hermes Agent. He broke it down as a solid OpenClaw alternative and…Continue reading on Medium »
- Your One-Click Private Agent Is a Chatbot Until You Make It Dangerous — Golden Jacket Field Notes, Week 14Continue reading on AWS Tip »
- Goose OSS — An open-source platform for seamless automation in engineering tasks, providing local execution and control. Customizable with any LLM for enhanced autonomy, enabling developers to efficiently handle...
- Khoj — Open-source app offers local or cloud AI chat with major LLMs, advanced semantic search, automation, personal newsletters, document indexing, image generation, and text-to-speech. Access from web,...
- 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...
Don't miss what's next. Subscribe to Openclaw Newsletter: