~ / blog / series / OpenClaw
❯ ls ~/blog/series/openclaw
11 posts
- partdatetitle
- 12026-03-05[AI Agent] Zero API Cost: Running OpenClaw on DGX Spark + Mac Mini
Full stack local AI agent: Mac Mini M4 as the always-on gateway, GX10 for inference, Telegram as the UI. No subscriptions, no cloud APIs. Six deployment lessons from the trenches.
- 22026-03-06[vLLM] Don't Add --enable-chunked-prefill to SSM Models
Adding --enable-chunked-prefill to a Qwen3.5-35B (SSM+MoE hybrid) dropped throughput from 47 tok/s to 5.7 tok/s. Why SSM recurrence and chunked prefill are fundamentally incompatible.
- 32026-03-07[vLLM] Ollama's KEEP_ALIVE Is Silently Eating Your vLLM Headroom
vLLM OOMed on restart despite 128GB unified memory. Cause: Ollama's KEEP_ALIVE=2h was holding 19-51GB in GPU. Diagnosis command, manual unload fix, and why to set KEEP_ALIVE=0 once vLLM is your primary stack.
- 42026-03-01[Benchmark] Pure MoE vs SSM Hybrid: Context Decay and Why It Matters for Agents
GLM-4.7-Flash hits 57.8 tok/s on short context but drops to 42 tok/s at 8K. Qwen3.5-35B SSM hybrid: 56 tok/s at short, 56 tok/s at 8K. Why agents with long system prompts should care about this difference.
- 52026-03-16[AI Agent] The Codex-Executor Pattern: Keeping Agent Sessions Small
Why we stopped having the OpenClaw agent orchestrate multi-step tasks directly, and started spawning Codex subprocesses instead. The pattern that keeps agent context minimal and tasks reliable.
- 62026-02-25[AI Agent] OpenClaw Config Hot-Reload: No Restart Needed
Spent weeks restarting the OpenClaw gateway for every config change. Then discovered the file watcher. What hot-reloads instantly, what still needs a restart, and how to tell auth failures from transient network errors.
- 72026-03-18[AI Agent] openclaw: When the Agent Calls for Help
How to wire a callhelp tool into a local agent loop so it can spawn Codex CLI mid-reasoning. One permission flag you must set, and why Claude's quota stays mine.
- 82026-03-19[AI Agent] openclaw: Why the Bot Went Silent — Tailscale, IPv6, and a Node.js Happy Eyeballs Trap
The bot process is running. The token is valid. Messages are being consumed. Nobody is home. A systematic takedown of every wrong hypothesis — and the hidden causal chain that connects Tailscale routing tables to silent sendMessage failures in Node.js.
- 92026-03-21[AI Agent] openclaw + 131K Context: When max_tokens Goes Negative
Connecting openclaw to a 131K context model and hitting 400 max_tokens must be at least 1, got -1292. The context budget math, the config key trap, and the fix.
- 102026-03-21[AI Agent] openclaw Real-Time Streaming via Telegram Bot API 9.5 sendMessageDraft
Replacing choppy editMessageText polling with Telegram's sendMessageDraft for live animated output. The patch, the think-block filter, and the optional chaining trap in DM chats.
- 112026-03-24[AI Agent] openclaw + ChatGPT OAuth: Run GPT-5.4 Agents Without API Credits
Your ChatGPT Plus subscription already includes GPT-5.4 with 1M context. openclaw's OAuth flow lets you use it for AI agents — zero API credits, one command. Full setup guide.