ShioriCode
Open-source alternative to Codex & Claude Code
145 followers
Open-source alternative to Codex & Claude Code
145 followers
A desktop interface for the coding-agent CLIs you already use — Codex, Claude Code, Cursor, Gemini, Kimi, and a hosted Shiori provider. Run long-running sessions with project-aware threads, stream agent activity into a readable timeline, and review generated diffs without leaving the app. Source-available, built for work that doesn't fit in one prompt.
Free
Launch Team
ArticosTest your messaging on your audience, right inside Claude or ChatGPT
Promoted
@samihindi Hey Sami, congrats on shipping 👋
The "desktop UI over agent CLIs you already use" angle is the right framing - agent CLIs are great at execution but terrible at session continuity once a task spans more than one coffee.
One question on the multi-provider story: when I switch between Claude Code and Codex inside the same ShioriCode project thread, what's shared and what isn't? Do they see the same diff stack and conversation history (re-injected as context), or is each provider its own isolated session that just happens to live in the same UI shell? Asking because the answer determines whether ShioriCode is a "launcher" or a real abstraction layer - very different products with very different moats.
The project-aware threads plus streaming diff review solves a real problem. Agent CLIs have terrible UX for work that spans hours across multiple files. Building RetainSure, we kept hitting context window limits in long-running sessions and had to externalize state manually. How does ShioriCode handle continuity when an underlying agent reaches its token limit mid-session?
@retain_dev We trigger the agent-native compaction once the token limit is reached :)
Hey Sami, went through ShioriCode's page and the "long sessions die in the terminal" line basically described my last week. one thing I wanted to ask, when you unify Claude Code, Codex, and Gemini in one workspace, are you normalizing the tool-call format under the hood or just rendering each one's native output? that abstraction call feels like make-or-break.
@axlerodd We are simply rendering each one's native output?
Could I access that from a remote machine? Run it on an Ubuntu and access it over X I suppose - or are there better ways?
@egor_kraev1 Not yet, we're still working on remote SSH access :)
Just wondering about the reason to build a desktop app over a CLI?
@sk_uxpin With a desktop app, it's easier to manage multiple agents. And clicking on a model is way faster than having to type in /model everytime you want to switch :)
@samihindi to be honest, clicking isn't faster to me - I'm using an external keyboard with my laptop and every time when I need to click or move the cursor, I need to take a hand off the keyboard, which isn't really fun :)
The 'long sessions die in the terminal' problem is real — I've lost entire Claude Code runs to scrollback with zero recovery. The branch-aware threading is the part that genuinely interests me, because right now switching between Claude Code and Codex mid-task means full context reset every time. One question: when you unify multiple providers in one thread, are you normalizing the diff format across them, or just rendering each CLI's native output? That distinction determines whether ShioriCode is a launcher or a real abstraction layer — very different products with very different moats