OpenClaw Newsletter - 2026-06-29
OpenClaw Newsletter
Monday, June 29, 2026Top Stories
- openclaw 2026.6.11-beta.2 — Highlights - More capable channel control: Slack relay mode, native Mattermost /oc_queue, and per-DM model overrides make channel operations easier to automate and tune. (#94707, #95546, #95120)...
- fix(exec-approval): stop misattributing Allow Always unavailability to policy — What Problem This Solves Fixes an issue where users running a command with a shell redirect (e.g. openclaw --version 2>&1) under ask=on-miss would see the approval prompt say "The effective approval...
- fix(/v1/responses): accept text field on requests for OpenAI SDK 6.x parity — Summary - Problem: /v1/responses rejects every request from clients using the current OpenAI Node SDK (openai >= 6.x) or @langchain/openai >= 1.x with Bad request - please check your parameters:...
- fix(process): add zombie reaper to prevent child process leak — Closes #97616 What Problem This Solves Long-running openclaw gateway processes in Docker accumulate zombie processes over time. One user reported 800+ zombie processes (openclaw-hooks, bash, codex,...
- feat(agents): per-agent model allowlist (with fallback to global) — When agents.list[].models is defined and non-empty, scope the visibility allowlist (used by the Control UI model picker, /model, and the gateway models.list catalog view) to that agent. Otherwise...
- fix(gateway): disable startup trace monitor on startup failure — Summary - add an explicit startupTrace.close() cleanup path in gateway startup tracing - call startupTrace.close() in the startup failure catch path so monitorEventLoopDelay is always disabled on...
- feat(plugins): add Control UI entry points — Summary - Problem: Plugins can expose gateway-authenticated HTTP routes, but there was no generic way for a plugin to register a first-class Control UI navigation entry. - What changed: Added generic...
- [codex] Further optimize Telegram media and reply context flow — Summary - add scoped indexes to the Telegram message cache so recentBefore and around avoid scanning and sorting the full cache - cache successful Telegram media resolutions by file identity and...
- fix: propagate gateway TLS fingerprints to bootstrap clients — Summary - Problem: Gateway clients bootstrapped outside callGateway (ACP, MCP channel bridge, and native approval handling) did not receive the local TLS certificate fingerprint for a self-signed...
- fix(google): add google-gemini-cli auth alias and expand model runtime bindings — What Problem This Solves close #95851 After re-onboarding from another provider (e.g. Claude CLI) to Gemini CLI via openclaw onboard, the chat becomes unresponsive — the UI shows infinite loading and...
- Prioritize foreground command queue work — Summary: - Add Low/Normal/High priority scheduling inside command lanes while preserving FIFO order within a priority. - Promote old low/normal entries after the starvation threshold so background...
- fix: add resilient fallback policy for user model overrides — Summary Describe the problem and fix in 2–5 bullets: - Problem: session/user model overrides always disabled configured fallback chains. - Why it matters: operators could configure provider fallback...
- [codex] fix native subagent completion retry cap — What Problem This Solves Fixes #97593. Codex native subagent completion delivery could retry forever when the parent handoff was permanently non-durable. The monitor marked the completion pending,...
- fix(cron): scope all-run job names — Summary Fixes #49175. cron.runs with scope: "all" still returns global run-log rows, but job-name enrichment now uses the same agent-scoped job visibility path as cron.list instead of reading every...
- fix(cron): warn when web_search allowlist has no provider — Closes #97654 AI-assisted change authored in Codex. What Problem This Solves Cron agentTurn jobs can explicitly set toolsAllow: ["web_search"] while no usable web search provider is selected or...
- fix(control-ui): preserve numeric chat input in composer — Add autocomplete/autocorrect/autocapitalize/spellcheck/inputmode attributes to chat textarea to prevent browser/OS numeric auto-formatting. Add regression tests for composer attributes and numeric...
- fix(logs): only report rotation when the log file actually shrank — PR title: fix(logs): only report rotation when the log file actually shrank Summary - Problem: openclaw logs --follow prints Log cursor reset (file rotated). to stderr whenever the gateway log grows...
- fix(security): report browser audit summary from plugin policy — Related: #86138, #76631, #96397, #92086, #92883 What Problem This Solves Fixes an issue where openclaw security audit could report browser control: enabled in the attack-surface summary when a...
- fix(googlechat): preserve thread reply target through delivery — Summary - Problem: Google Chat replies could leave the inbound Chat thread and appear as new top-level messages. - Why it matters: delayed or tool-driven assistant replies break group-thread UX when...
- fix(memory): record structured dreaming outcomes — Closes #97690 AI-assisted: yes, authored with Codex. What Problem This Solves Fixes an issue where users and operators reading memory dreaming event logs could see artifact-write records but could...
- [Bug]: Subagent completion event is marked as delivered upon queueing without ACK, causing sessions_yield parent to hang — Bug type Regression (worked before, now fails) Beta release blocker No Summary Description In the current OpenClaw Subagent architecture, when a child agent completes a task, the registry layer...
- fix(agents): warm context-window cache for the gateway daemon — What The long-running gateway daemon never warms the context-window cache, so synchronous lookups on the status path return DEFAULT_CONTEXT_TOKENS = 200_000 until the first agent run completes. That...
- fix(agents): load symlinked workspace bootstrap files — Summary - Load explicit fixed workspace bootstrap files such as AGENTS.md, SOUL.md, USER.md, and MEMORY.md when they are symlinks to readable regular files outside the workspace. - Keep the fallback...
- fix: gate same-model retries to idempotent calls — Summary This is a narrow ProjectClownfish replacement for the same-model transient retry behavior proposed in #44385. It preserves attribution to @nicolasgrasset and source PR...
- fix(exec-policy): accept trusted ~/.openclaw symlink at home boundary — Fixes #72572. Problem The exec-approvals symlink hardening from #72377 only relaxed the restriction on the OPENCLAW_HOME root itself, so a symlinked ~/.openclaw immediate child of a real home...
- fix(discord): truncate forum thread names on a surrogate boundary — What Problem This Solves Fixes an issue where Discord forum/media channel thread names can come out with a broken character (a � or mangled emoji) when the derived name hits Discord's 100-character...
- test(auth): add oauth e2e regression coverage — Summary - add an OAuth auth-regressions e2e suite covering CLI and gateway auth-status surfaces - add targeted regression tests for stale auth order, shared refresh coordination, unhealthy status...
- feat(agents): expose tools and their schemas via new after_tools_resolved hook [AI-assisted] — Summary - Problem: Plugins have no way to discover which tools are available to an agent or inspect their input schemas. The plugin API exposes registerTool() but no listTools() or getTools()....
- Fix voice-call streaming provider resolution — Fixes #97738 What Problem This Solves talk.catalog reports transcription.activeProvider from the Talk/voice-call streaming config, but realtime transcription provider listing could return only the...
- fix(matrix): reverse-proxy path support for Matrix homeserver URLs — Summary Fixes Matrix transport URL construction for homeservers behind reverse proxies with path prefixes. The apiPrefix public config surface was removed following ClawSweeper review — the fix now...
- OpenClaw repo stats for 2026-06-29 — 380931 stars, 79801 forks, 2708 contributors
- OpenClaw sponsors (110 total) — 110 sponsors supporting OpenClaw
- fix(workers): harden ClawHub background retries (#2924)
- fix: restore catalog feed schema v1 (#2909)
- openclaw/openclaw on Docker Hub — Official openclaw/openclaw docker images.
- sulpikar2/Qwen3.5-4B-Claude-4.6-Opus-Reasoning-openclaw
- 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...
- Anthropic says OpenClaw-style Claude CLI usage is allowed again
- Polling Agents in AI Assistants: 11 Implementation Patterns — Polling agents are one of the least glamorous parts of AI assistant architecture, but they are also...
- What Is the A2A Protocol? Agent Cards and Tasks Explained — The A2A Protocol, short for Agent2Agent Protocol, is an open standard for communication between...
- Never Miss a Beat: Cron Scheduling in OpenClaw — A practical tutorial from the sol.ai blog.
- Build a Private Photo Assistant on Telegram with OpenClaw + MiniCPM-V 4.6 — Most AI assistants can read text, write code, and automate workflows. But the moment a user sends a...
- QA Logger — OpenClaw Plugin to Turn AI Chats into a Searchable Diary — QA Logger: Turn Every AI Chat Into a Searchable Diary\n\nThe Problem\n\nOpenClaw sessions...
- Your OpenClaw Bill Is Bleeding Tokens. Here’s What We Measured — and How to Fix It — Memory bloat, compaction loss, and a retrieval-first path: ~32% less token spend on the AppWorld dev...
- Next-Iteration Improvements: Optimizing Personal Agentic AI Assistant with Llama.cpp, Gemma 4 12B and MCP — Background Building a $0 personal agentic AI assistant means you don't have the luxury of...
- Your Personal AI Assistant: Automating Gmail Insights with OpenClaw — Imagine starting every morning with a concise briefing of your latest emails and a personalized list of topics worth learning — without…Continue reading on Medium »
- NanoClaw — Lightweight AI assistant runs in Apple containers with isolated Linux container processes for superior OS-level security, transparency, and privacy compared to complex chatbots, enabling quick code...
- Odysseus — Self-hosted AI workspace offering chat, tool use, and autonomous agent support. Enables email and research tasks securely on user hardware, keeps all data local, requires no external servers, and...
- Private GPT — Ask questions to your documents without an internet connection, using the power of LLMs. 100% private, no data leaves your execution environment at any point. You can ingest documents and ask...
- Reins — Empowering LLM researchers and hobbyists with seamless control over self-hosted models. Connect remotely, customize prompts, manage chats, and fine-tune configurations. All in one intuitive app.
- my openclaw analyst flagged $ALLINDOGE as a buy. noted 🦞 [SCAN] $ALLINDOGE [MC] $4.0M [VOL] $36.7K [LIQ] $217.1
- openclaw called it on $$NRG again. the model doesn't miss $$NRG — quick check • Current MC: $2.7M • 24h Volume: $14.
- openclaw been right on $FRAG twice already. trusting the scan $FRAG MC $911.9K → $11.7M (+1179%) 24h vol $552.9K W
- just checked $BOBO —openclaw showing strong upside signals > BOBO MKT CAP $7.9M 24H VOL $123.7K LIQUID
- openclaw confirmed what the chart was already telling me on $MASK [SCAN] $MASK [MC] $1.4M [VOL] $58.1K [LIQ] $2
- openclaw on $PRELODE —momentum strong, risk low. easy $PRELODE · health 🟥⬜⬜⬜⬜⬜⬜⬜⬜⬜ mc $8.4M · vol $53.9K · liq $219.8
- been tracking $BULL on openclaw for a week. forecast holding strong $BULL ┃ mc $2.0M ┃ vol $141.4K ┃ liq $243.1K ┃ tgt
- $BP looking like a clean entry according to openclaw $BP ┃ mc $1.8M ┃ vol $85.8K ┃ liq $224.0K ┃ tgt $7.1M ┃ +291% W
- openclaw flagged $gork as a strong hold. not surprised $gork ▁▂▃▅▇ → $10.7M mc $3.4M · vol $17.9K · liq $248.3K · +21
- @grok @HumbertoGatolo @grok Contestale anda que no parece entender que los ordenadores consumen electricidad o quizas no
- openclaw on $RNT —momentum strong, risk low. easy $RNT ├ mc $605.4K ├ vol $17.6K ├ liq $266.7K └ tgt $1.5M (×2.6)
- openclaw on $CODEC —signals all pointing up $CODEC MC $2.9M → $11.0M (+279%) 24h vol $58.8K Was Made Using - @Open
- just ran $WhiteWhale through openclaw. risk is low, upside is real $WhiteWhale ▁▂▃▅▇ → $21.4M mc $3.2M · vol $785.5K
- $SIGMA momentum score just came in hot on openclaw 🔥 $SIGMA ├ mc $1.3M ├ vol $71.2K ├ liq $313.6K └ tgt $4.1M (×3.
- openclaw confirmed what the chart was already telling me on $67 $67 $2.3M ▰▰▱▱▱▱▱▱▱▱ $12.2M vol $256.8K · liq $348.6
- openclaw called $FARTBOY bullish. i'm listening $FARTBOY MC $9.5M → $24.9M (+163%) 24h vol $12.7K Was Made Using -
- $UwU looking like a clean entry according to openclaw > UWU MKT CAP $42.9M 24H VOL $627.2K LIQUIDITY $
- $ZERO looking healthy according to openclaw. adding more $ZERO current → $1.9M target → $11.1M vol 24h → $69
- openclaw on $retire —signals all pointing up $retire MC $2.9M → $8.1M (+184%) 24h vol $68.8K Was Made Using - @Ope
- $CHILLHOUSE is looking good on openclaw right now ngl $CHILLHOUSE ┃ mc $3.5M ┃ vol $378.3K ┃ liq $540.7K ┃ tgt $18.8M
Trending on X
- my openclaw analyst flagged $ALLINDOGE as a buy. noted 🦞 [SCAN] $ALLINDOGE [MC] $4.0M [VOL] $36.7K [LIQ] $217.1
- openclaw called it on $$NRG again. the model doesn't miss $$NRG — quick check • Current MC: $2.7M • 24h Volume: $14.
- openclaw been right on $FRAG twice already. trusting the scan $FRAG MC $911.9K → $11.7M (+1179%) 24h vol $552.9K W
- just checked $BOBO —openclaw showing strong upside signals > BOBO MKT CAP $7.9M 24H VOL $123.7K LIQUID
- openclaw confirmed what the chart was already telling me on $MASK [SCAN] $MASK [MC] $1.4M [VOL] $58.1K [LIQ] $2
- openclaw on $PRELODE —momentum strong, risk low. easy $PRELODE · health 🟥⬜⬜⬜⬜⬜⬜⬜⬜⬜ mc $8.4M · vol $53.9K · liq $219.8
- been tracking $BULL on openclaw for a week. forecast holding strong $BULL ┃ mc $2.0M ┃ vol $141.4K ┃ liq $243.1K ┃ tgt
- $BP looking like a clean entry according to openclaw $BP ┃ mc $1.8M ┃ vol $85.8K ┃ liq $224.0K ┃ tgt $7.1M ┃ +291% W
- openclaw flagged $gork as a strong hold. not surprised $gork ▁▂▃▅▇ → $10.7M mc $3.4M · vol $17.9K · liq $248.3K · +21
- @grok @HumbertoGatolo @grok Contestale anda que no parece entender que los ordenadores consumen electricidad o quizas no
- openclaw on $RNT —momentum strong, risk low. easy $RNT ├ mc $605.4K ├ vol $17.6K ├ liq $266.7K └ tgt $1.5M (×2.6)
- openclaw on $CODEC —signals all pointing up $CODEC MC $2.9M → $11.0M (+279%) 24h vol $58.8K Was Made Using - @Open
- just ran $WhiteWhale through openclaw. risk is low, upside is real $WhiteWhale ▁▂▃▅▇ → $21.4M mc $3.2M · vol $785.5K
- $SIGMA momentum score just came in hot on openclaw 🔥 $SIGMA ├ mc $1.3M ├ vol $71.2K ├ liq $313.6K └ tgt $4.1M (×3.
- openclaw confirmed what the chart was already telling me on $67 $67 $2.3M ▰▰▱▱▱▱▱▱▱▱ $12.2M vol $256.8K · liq $348.6
- openclaw called $FARTBOY bullish. i'm listening $FARTBOY MC $9.5M → $24.9M (+163%) 24h vol $12.7K Was Made Using -
- $UwU looking like a clean entry according to openclaw > UWU MKT CAP $42.9M 24H VOL $627.2K LIQUIDITY $
- $ZERO looking healthy according to openclaw. adding more $ZERO current → $1.9M target → $11.1M vol 24h → $69
- openclaw on $retire —signals all pointing up $retire MC $2.9M → $8.1M (+184%) 24h vol $68.8K Was Made Using - @Ope
- $CHILLHOUSE is looking good on openclaw right now ngl $CHILLHOUSE ┃ mc $3.5M ┃ vol $378.3K ┃ liq $540.7K ┃ tgt $18.8M
Releases
- openclaw 2026.6.11-beta.2 — Highlights - More capable channel control: Slack relay mode, native Mattermost /oc_queue, and per-DM model overrides make channel operations easier to automate and tune. (#94707, #95546, #95120)...
- OpenClaw repo stats for 2026-06-29 — 380931 stars, 79801 forks, 2708 contributors
- openclaw/openclaw on Docker Hub — Official openclaw/openclaw docker images.
Community
- fix(exec-approval): stop misattributing Allow Always unavailability to policy — What Problem This Solves Fixes an issue where users running a command with a shell redirect (e.g. openclaw --version 2>&1) under ask=on-miss would see the approval prompt say "The effective approval...
- fix(/v1/responses): accept text field on requests for OpenAI SDK 6.x parity — Summary - Problem: /v1/responses rejects every request from clients using the current OpenAI Node SDK (openai >= 6.x) or @langchain/openai >= 1.x with Bad request - please check your parameters:...
- fix(process): add zombie reaper to prevent child process leak — Closes #97616 What Problem This Solves Long-running openclaw gateway processes in Docker accumulate zombie processes over time. One user reported 800+ zombie processes (openclaw-hooks, bash, codex,...
- feat(agents): per-agent model allowlist (with fallback to global) — When agents.list[].models is defined and non-empty, scope the visibility allowlist (used by the Control UI model picker, /model, and the gateway models.list catalog view) to that agent. Otherwise...
- fix(gateway): disable startup trace monitor on startup failure — Summary - add an explicit startupTrace.close() cleanup path in gateway startup tracing - call startupTrace.close() in the startup failure catch path so monitorEventLoopDelay is always disabled on...
- feat(plugins): add Control UI entry points — Summary - Problem: Plugins can expose gateway-authenticated HTTP routes, but there was no generic way for a plugin to register a first-class Control UI navigation entry. - What changed: Added generic...
- [codex] Further optimize Telegram media and reply context flow — Summary - add scoped indexes to the Telegram message cache so recentBefore and around avoid scanning and sorting the full cache - cache successful Telegram media resolutions by file identity and...
- fix: propagate gateway TLS fingerprints to bootstrap clients — Summary - Problem: Gateway clients bootstrapped outside callGateway (ACP, MCP channel bridge, and native approval handling) did not receive the local TLS certificate fingerprint for a self-signed...
- fix(google): add google-gemini-cli auth alias and expand model runtime bindings — What Problem This Solves close #95851 After re-onboarding from another provider (e.g. Claude CLI) to Gemini CLI via openclaw onboard, the chat becomes unresponsive — the UI shows infinite loading and...
- Prioritize foreground command queue work — Summary: - Add Low/Normal/High priority scheduling inside command lanes while preserving FIFO order within a priority. - Promote old low/normal entries after the starvation threshold so background...
- fix: add resilient fallback policy for user model overrides — Summary Describe the problem and fix in 2–5 bullets: - Problem: session/user model overrides always disabled configured fallback chains. - Why it matters: operators could configure provider fallback...
- [codex] fix native subagent completion retry cap — What Problem This Solves Fixes #97593. Codex native subagent completion delivery could retry forever when the parent handoff was permanently non-durable. The monitor marked the completion pending,...
- fix(cron): scope all-run job names — Summary Fixes #49175. cron.runs with scope: "all" still returns global run-log rows, but job-name enrichment now uses the same agent-scoped job visibility path as cron.list instead of reading every...
- fix(cron): warn when web_search allowlist has no provider — Closes #97654 AI-assisted change authored in Codex. What Problem This Solves Cron agentTurn jobs can explicitly set toolsAllow: ["web_search"] while no usable web search provider is selected or...
- fix(control-ui): preserve numeric chat input in composer — Add autocomplete/autocorrect/autocapitalize/spellcheck/inputmode attributes to chat textarea to prevent browser/OS numeric auto-formatting. Add regression tests for composer attributes and numeric...
- fix(logs): only report rotation when the log file actually shrank — PR title: fix(logs): only report rotation when the log file actually shrank Summary - Problem: openclaw logs --follow prints Log cursor reset (file rotated). to stderr whenever the gateway log grows...
- fix(security): report browser audit summary from plugin policy — Related: #86138, #76631, #96397, #92086, #92883 What Problem This Solves Fixes an issue where openclaw security audit could report browser control: enabled in the attack-surface summary when a...
- fix(googlechat): preserve thread reply target through delivery — Summary - Problem: Google Chat replies could leave the inbound Chat thread and appear as new top-level messages. - Why it matters: delayed or tool-driven assistant replies break group-thread UX when...
- fix(memory): record structured dreaming outcomes — Closes #97690 AI-assisted: yes, authored with Codex. What Problem This Solves Fixes an issue where users and operators reading memory dreaming event logs could see artifact-write records but could...
- [Bug]: Subagent completion event is marked as delivered upon queueing without ACK, causing sessions_yield parent to hang — Bug type Regression (worked before, now fails) Beta release blocker No Summary Description In the current OpenClaw Subagent architecture, when a child agent completes a task, the registry layer...
- fix(agents): warm context-window cache for the gateway daemon — What The long-running gateway daemon never warms the context-window cache, so synchronous lookups on the status path return DEFAULT_CONTEXT_TOKENS = 200_000 until the first agent run completes. That...
- fix(agents): load symlinked workspace bootstrap files — Summary - Load explicit fixed workspace bootstrap files such as AGENTS.md, SOUL.md, USER.md, and MEMORY.md when they are symlinks to readable regular files outside the workspace. - Keep the fallback...
- fix: gate same-model retries to idempotent calls — Summary This is a narrow ProjectClownfish replacement for the same-model transient retry behavior proposed in #44385. It preserves attribution to @nicolasgrasset and source PR...
- fix(exec-policy): accept trusted ~/.openclaw symlink at home boundary — Fixes #72572. Problem The exec-approvals symlink hardening from #72377 only relaxed the restriction on the OPENCLAW_HOME root itself, so a symlinked ~/.openclaw immediate child of a real home...
- fix(discord): truncate forum thread names on a surrogate boundary — What Problem This Solves Fixes an issue where Discord forum/media channel thread names can come out with a broken character (a � or mangled emoji) when the derived name hits Discord's 100-character...
- test(auth): add oauth e2e regression coverage — Summary - add an OAuth auth-regressions e2e suite covering CLI and gateway auth-status surfaces - add targeted regression tests for stale auth order, shared refresh coordination, unhealthy status...
- feat(agents): expose tools and their schemas via new after_tools_resolved hook [AI-assisted] — Summary - Problem: Plugins have no way to discover which tools are available to an agent or inspect their input schemas. The plugin API exposes registerTool() but no listTools() or getTools()....
- Fix voice-call streaming provider resolution — Fixes #97738 What Problem This Solves talk.catalog reports transcription.activeProvider from the Talk/voice-call streaming config, but realtime transcription provider listing could return only the...
- fix(matrix): reverse-proxy path support for Matrix homeserver URLs — Summary Fixes Matrix transport URL construction for homeservers behind reverse proxies with path prefixes. The apiPrefix public config surface was removed following ClawSweeper review — the fix now...
- OpenClaw sponsors (110 total) — 110 sponsors supporting OpenClaw
- fix(workers): harden ClawHub background retries (#2924)
- fix: restore catalog feed schema v1 (#2909)
News
- sulpikar2/Qwen3.5-4B-Claude-4.6-Opus-Reasoning-openclaw
- 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...
- Anthropic says OpenClaw-style Claude CLI usage is allowed again
- Polling Agents in AI Assistants: 11 Implementation Patterns — Polling agents are one of the least glamorous parts of AI assistant architecture, but they are also...
- What Is the A2A Protocol? Agent Cards and Tasks Explained — The A2A Protocol, short for Agent2Agent Protocol, is an open standard for communication between...
- Never Miss a Beat: Cron Scheduling in OpenClaw — A practical tutorial from the sol.ai blog.
- Build a Private Photo Assistant on Telegram with OpenClaw + MiniCPM-V 4.6 — Most AI assistants can read text, write code, and automate workflows. But the moment a user sends a...
- QA Logger — OpenClaw Plugin to Turn AI Chats into a Searchable Diary — QA Logger: Turn Every AI Chat Into a Searchable Diary\n\nThe Problem\n\nOpenClaw sessions...
- Your OpenClaw Bill Is Bleeding Tokens. Here’s What We Measured — and How to Fix It — Memory bloat, compaction loss, and a retrieval-first path: ~32% less token spend on the AppWorld dev...
- Next-Iteration Improvements: Optimizing Personal Agentic AI Assistant with Llama.cpp, Gemma 4 12B and MCP — Background Building a $0 personal agentic AI assistant means you don't have the luxury of...
- Your Personal AI Assistant: Automating Gmail Insights with OpenClaw — Imagine starting every morning with a concise briefing of your latest emails and a personalized list of topics worth learning — without…Continue reading on Medium »
- NanoClaw — Lightweight AI assistant runs in Apple containers with isolated Linux container processes for superior OS-level security, transparency, and privacy compared to complex chatbots, enabling quick code...
- Odysseus — Self-hosted AI workspace offering chat, tool use, and autonomous agent support. Enables email and research tasks securely on user hardware, keeps all data local, requires no external servers, and...
- Private GPT — Ask questions to your documents without an internet connection, using the power of LLMs. 100% private, no data leaves your execution environment at any point. You can ingest documents and ask...
- Reins — Empowering LLM researchers and hobbyists with seamless control over self-hosted models. Connect remotely, customize prompts, manage chats, and fine-tune configurations. All in one intuitive app.
Don't miss what's next. Subscribe to Openclaw Newsletter: