David Parys

Starting June 15, paid Claude plans can claim a dedicated monthly credit for programmatic usage. The credit covers usage of: - Claude Agent SDK - claude -p - Claude Code GitHub Actions - Third-party apps built on the Agent SDK

This 3-minute video from Anthropic will make you realize you've been using Claude wrong this whole time "when you open Claude Code without a CLAUDE.md file, it has to start fresh every single time"

Just a couple of prompts from start to finish into a playable game! Pixel-perfect art > Codex App w/ GPT 5.5 High > Images 2.0 for animations (except walkcycles) > WAN 2.0 for walkcycles (cheap!) > 11Labs for bgm/sfx (in 1 prompt) > Phaser 4 Sound ON🔉 Full tutorial anyone?

Over the last few days, I explored 4 AI agent orchestrator apps: - DP Code - Emdash - Letta Code - Superset Most of them position themselves as a "harness aggregator" or an Agentic Development Environment, a layer on top of your existing AI tools that tries to unify workflows and coordinate multiple agents. I connected my Claude Code, Codex, Cursor, Minimax API, and OpenCode across these platforms and spent a few hours with each to understand how they actually perform in real usage. Here’s what I found: DP Code > Started off as a T3 Chat–style clone, but has gone through multiple revamps, currently in alpha and now looks very close to the Codex app > Lets you connect Gemini, Claude, OpenAI, and OpenCode > You can split work across multiple chats and run multiple agents in parallel > Switching providers mid-task is supported, which is a really nice touch Emdash > YC-backed and focused on running multiple coding agents in parallel > Supports 20+ AI providers in a single project with separate Git workflows > Has UI-based plugin, MCP, and skills installation On Windows it felt a bit sluggish under the hood it was opening separate terminals per agent and lagged a bit (might be smoother on macOS) Letta Code > A memory-first coding agent similar to hermes agent i guess > Automatically detects already logged-in AI providers and lets you use them directly which didn't work with others > Feels quite similar to OpenCode, but with stronger memory handling > Also supports cron jobs, which is useful for automation workflows Superset > Very similar to Emdash (or vice versa) in terms of positioning > Currently macOS-only, I tried running it via an Electron wrapper, but it didn’t work well > Focuses on monitoring multiple agents from a single interface > Each agent’s changes are isolated, which is helpful for managing parallel workflows Overall, all of them are pushing in the same direction of multi-agent workflows and provider abstraction, it will be interesting to see which stays in the long term game.

Meet Chief. The first inbox with good judgment.