Clinch – Local-first Warp fork built for agent session management
Warp was my daily driver for years, and I still love the core product. But I wanted more privacy, less Oz agent stuff and a better UX experience for managing all my Claude Code/Codex sessions, especially across the repos I'm working on at once. So I forked Warp, gutted all the telemetry, account requirements, Oz agent stuff and have fine tuned the UX over the months to optimize it for session management. The awesome Warp terminal is still first class (auto complete, auto command suggestion etc) but with lots of QoL improvements for agent session management and organization. Some notable…
What it does
Manage all your Claude Code and Codex sessions in one terminal built for agents, with automatic resume, quick actions, and one-click session transfers.
Every session auto-resumes on restart. Self-adapting UI. Quick-response and action buttons. One-click session transfers between Claude Code and Codex. Install Clinch Terminal from https://github.com/elliot-ylambda/clinch-terminal. Follow the official install instructions, review the installer source, and verify the release before installing. Unnotarized public preview for macOS 14+. The installer authenticates a signed release, verifies its digest, and stages Clinch in Applications. Read the installer source · View on GitHub → Send a line from your phone. It runs in the session already open on your Mac. I run Claude Code and Codex all day, in a lot of tabs. Every feature below started with…from clinch.sh
In the maker’s words, at launch
Warp was my daily driver for years, and I still love the core product. But I wanted more privacy, less Oz agent stuff and a better UX experience for managing all my Claude Code/Codex sessions, especially across the repos I'm working on at once. So I forked Warp, gutted all the telemetry, account requirements, Oz agent stuff and have fine tuned the UX over the months to optimize it for session management. The awesome Warp terminal is still first class (auto complete, auto command suggestion etc) but with lots of QoL improvements for agent session management and organization. Some notable features: * Claude Code and Codex sessions auto resume with the correct flags if you ever quit/restart Clinch -- this is a huge issue in Warp for me and always prevented me from updating to the latest version of Warp. * Clinch CLI + Skill so Claude and Codex can auto create new tabs for you and even auto suggest quick response buttons (see below). This is especially nice when you ask Claude to run a local web server, it will auto run it in a new tab so it's not hidden in the background of a Claude session. * Remote control using Tailscale, use and control Clinch from your phone. * Left side panel contains all your sessions/terminal tabs (switch between with CMD + SHIFT + [) for a repo. Each repo/project has its own tab in the header, quickly switch between them with CMD + [. * Instantly see agent status (working, done, needs input, idle). * Customizable sections in the left side panel to organize your sessions into groups i.e. "In Progress" and "Backlog" etc. * Customizable quick response buttons in the footer of Claude and Codex sessions to quickly respond with: "/compact", "LGTM, continue" or "Review this with Fable" etc. * Action buttons in the footer: One-click transfer a Claude Code session to a Codex session and vice versa. One-click fork session in new tab. 100% open source and free. No account or signup needed. This is a personal project that I enjoy working on in my spare time. macOS only for now, but if there is interest I will add Linux support next.
Does the same job
all alternatives →
HarnessRouter Community Edition21d ago · harnessrouter.ai · ▲392Open-source unified interface for agent harnesses
More ai this month
the category →
I trained a 125M-parameter transformer to autocomplete piano performances in real time (~108 notes/sec on an iPhone 15). The idea is basically GitHub Copilot or Tabnine, except instead of prompting it with code, you prompt it by playing a few notes on a MIDI piano. The model then continues what you played, entirely on-device. The app is free if anyone wants to try it. Happy to answer questions about the model, training, Core ML, or the many things that didn't work.
AI · 17d ago · simedw.com
Astute▲585Automate your B2B brand going viral, with new media creators
AI · 18d ago · company-app.joinastute.com
Hey HN, Henry from Cactus here! We previously released Cactus Needle, a 14MB agentic LLM for tool call, device use, and structured extraction for phones, wearables, smart homes, small robots and microcontrollers. We got really great feedback here, and have now incorporated the suggestions to release Needle 2. The whole model is a single 14MB binary that runs a full session in 28MB of RAM; 45m parameters at 2bit compression. Needle hits 500 tokens/sec decode speed on a Raspberry Pi 5, sits between 400-1,500 tokens/sec on VR devices like Meta Quest 3S and Apple Vision Pro, and ranges…
AI · 27d ago · cactuscompute.com
Launched alongside, August 2026
the whole month →- TL
Life & fun · 10d ago · louisabraham.github.io
- SA
Hello HN! I found that picking out plausible but diverse skin tones for my digital art and game development projects was kind of difficult, and I got curious about if there was a way to define a color space that made it easy. I've built a color picker and procedural generation algorithm based on the space as well as a bunch of other fun js features and demos throughout the page that use the equations. If you find it interesting, I have lots of explanations of how I built it and what properties the space has. The methodology might be a bit shaky, but hopefully the result is as helpful for…
Life & fun · Aug 2026 · toneyalexander.github.io
I trained a 125M-parameter transformer to autocomplete piano performances in real time (~108 notes/sec on an iPhone 15). The idea is basically GitHub Copilot or Tabnine, except instead of prompting it with code, you prompt it by playing a few notes on a MIDI piano. The model then continues what you played, entirely on-device. The app is free if anyone wants to try it. Happy to answer questions about the model, training, Core ML, or the many things that didn't work.
AI · 17d ago · simedw.com