Fetching from the wire…
Public story · 2026-09-08 · high
Codenotch pulls Claude Code, Cursor, Codex and three other tools' local session data instead of asking you to log in again.
Why now: Version 1.6.0 went up September 8, a day after 1.5.0, with the repo already past 1,000 stars since its September 5 launch.
Codenotch shows how close you are to usage limits across seven AI coding tools, without another login. That matters if you pay for more than one assistant. Instead of checking Claude Code, Cursor and Codex separately, one menu bar icon shows every limit before a session cuts off mid-task.
Codenotch is a Swift menu-bar app that pulls this from each tool's own local session. It reads Claude Code's /usage output directly, falling back to the OAuth token in your login keychain when that's unavailable. For Cursor it reads the local SQLite session state. For Codex it hits the ChatGPT usage endpoint for 5-hour and weekly limits. Antigravity, GLM through Z.ai and Grok's ~/.grok/auth.json round out the list.
Codenotch works by borrowing credentials already on disk, in keychains, SQLite files and auth JSON, rather than asking for a fresh login. That access covers session tokens for six services the moment you install it, a broader grant than a web dashboard that only sees what you paste in.
The repo has 1,011 stars since its September 5 launch. Version 1.6.0 shipped September 8, a day after 1.5.0, and a Rust and Tauri port for Windows lives in the same repo.
Each link below shares sources, entities, or timing with this story.
It wires ChatGPT/Codex and Grok into DeepSeek Harness by OAuth in the web UI, and imports Claude credentials directly from an existing Claude Code session via the macOS Keychain or ~/.claude/.credentials.json. Tokens land in ~/.dsh/plugins/subscriptions/auth.json at mode 0600...
akitaonrails/ai-memory (1,229 stars, +184 this week, MIT, Rust, 900 commits) inverts the usual design: hooks emit sanitized lifecycle observations during a session, and the server compiles them into coherent wiki pages at session end or PreCompact, "compiled from observations...
Three things happened this month that only make sense together. Agent Plugins 1.0 shipped co-signed by six competitors: AWS, Anysphere, Microsoft, OpenAI, Vercel and Google (GitHub Changelog). It makes skills-plus-MCP bundles portable across clients. OpenAI's August 11 Codex c...
wanshuiyin/HERO-Anti-OverDefense went from creation to 68 stars in a single day. HERO is Hashing, Edge cases, Rubrics, Overbuild, and the claim is that agent over-engineering isn't diffuse but falls into four recognizable shapes suppressible with a portable prompt contract acr...
Warp released its client codebase under AGPL-3.0, surged to 56,000 GitHub stars and #2 on GitHub Trending. But the real story isn't the open-sourcing. It's the repositioning. Warp isn't calling itself a terminal anymore. It's an "agentic development environment." The product n...
claude-mem hit 80,189 stars at v12.6.4, with 1,840 commits and 109 contributors. It hooks five agent lifecycle events to capture observations, compresses them through Claude's agent SDK into SQLite, and reinjects relevant context on new sessions. No manual tagging. One npx com...
MindPattern daily
One email a day at 7 AM. Sources and a take on every story. Unsubscribe anytime.