Blog

Latest news, tutorials, and updates from Octomind

RSS feed
The Octomind Cloud import screen listing what comes over from an existing agent: persona, memories, skills and scheduled jobs, with API keys marked as not imported

Bring Your Agent to Octomind Cloud: From ChatGPT, Claude Code, Codex, OpenClaw or Hermes

Export ChatGPT memory, or bring Claude Code, Codex, OpenClaw or Hermes memories, skills and schedules to Octomind Cloud in one step. Plus new media models.

October 15, 2026 Don Karter Don Karter
A coding agent answering in a phone chat while the laptop that would normally run it stays closed, with the work running on a cloud machine

Claude Code Remote Control Without the Laptop: Code From Your Phone

Claude Code Remote Control needs your laptop on and awake. What it does well, where it stops, and how to code from your phone with the lid closed.

October 15, 2026 Don Karter Don Karter
Octomind 0.58.0: one decision-model configuration alongside session supervision costs and verification outcomes

Octomind 0.58.0: Configure Quick Decisions in One Place and See What They Cost

Octomind 0.58.0 centralizes decision-model settings, tracks supervision costs per session, and keeps verification tied to the output actually seen.

October 14, 2026 Don Karter Don Karter
Harness engineering explained with one experiment: the same model in two harnesses, one stopping at "Fixed" with a failing test, the other sent back by a test validator until both tests pass

What Is Harness Engineering? Why the System Around the Model Decides What Ships

Harness engineering is designing the checks and loops around a coding agent. Same model, same bug: the harness alone decided whether the fix shipped.

October 14, 2026 Don Karter Don Karter
An LLM wiki of agent-maintained markdown pages next to an octobrain memory store, both fed the same three source documents

LLM Wiki: Let Your Agent Keep Its Own Notes (and When Real Memory Wins)

Karpathy's LLM wiki lets an agent maintain its own markdown notes. I ran one against octobrain memory on the same sources. Here's where each one wins.

October 13, 2026 Don Karter Don Karter
Playwright CLI, Playwright MCP and Chrome DevTools MCP compared by tool-schema tokens and snapshot size on the same Hacker News page

Playwright CLI vs MCP: give your agent a browser that actually works

I ran Playwright CLI, Playwright MCP and Chrome DevTools MCP on the same page. The snapshots came out nearly identical; the real cost sits elsewhere.

October 12, 2026 Don Karter Don Karter
MCP vs CLI vs skills compared by context cost: MCP tool schemas, CLI help text and SKILL.md files measured on Octomind servers

MCP vs CLI vs Skills: When a Shell Command Is Enough and When You Need a Server

MCP vs CLI vs skills, measured on our own servers: 14 MCP tools cost 7,553 tokens before a prompt. When a shell command wins and when a server earns it.

October 11, 2026 Don Karter Don Karter
Benchmark scoreboard comparing open weight models GLM-5.3 and GLM-5.3 Flash with closed weight Claude Opus and GPT models on 20 real pull requests, solved count and total cost

Open Weight Models That Actually Hold Up on Real Coding Work

We ran open weight models on real pull requests. GLM-5.3 Flash landed 19 of 20 fixes, same as Opus in Claude Code, for $1.64 instead of $41.34.

October 10, 2026 Don Karter Don Karter
Five evaluation criteria for choosing an AI coding agent based on repository tasks and verified results

How to choose an AI coding agent for real work

Choose an AI coding agent with repo-level tests for autonomy, context, verification, integrations, and cost. Judge working changes, not polished demos.

October 9, 2026 Don Karter Don Karter
Octomind 0.57.0: an assistant reply quoted into the terminal prompt with the answer typed underneath

Quote the Part That Matters: Octomind 0.57.0 Answers Replies Point by Point

Octomind 0.57.0 quotes an earlier reply into your next prompt so you can answer it point by point. Plus verification, vision, and long-session fixes.

October 8, 2026 Don Karter Don Karter
An AI agent connects through a Postgres MCP server to a database with separate read and write permissions

Postgres MCP: connect your agent to real data

Connect your agent to real data with Postgres MCP: choose a server, configure credentials, register tools, and verify queries before adding write access.

October 8, 2026 Don Karter Don Karter
A comparison of LiteLLM and OpenRouter gateway costs, controls, and setup

LiteLLM vs OpenRouter: Cost, Control, and First Config (2026)

OpenRouter vs LiteLLM comes down to gateway fees, who operates the proxy, and where provider keys live. Compare the math, controls, and first config.

October 7, 2026 Don Karter Don Karter