The open-source community is INSANE 🤯 Claude Code routines dropped literal hours ago, and this guy's already shipped a clone. Meet Multica Autopilot: Now you can run routines entirely locally using whatever agent you want (Opencode, Codex, Hermes, OpenClaw etc). repo in 🧵↓

Repo → Shoutout to @jiayuan_jy for building this. Don't forget to drop a star ⭐️!

Oh niiiice, looks like they just launched a website and an X account too! X → Website →

Open-source is actually unstoppable xd Claude drops routines → Multica already ships local version with any agent. Running it now for Polymarket monitoring. This changes the degen game completely. LFG

the real value isn't just local vs cloud. it's that CLAUDE.md and your architecture rules actually get enforced consistently. ship a feature, it follows your standards. break a pattern, routine catches it. that's the difference between code that drifts and code that stays maintainable

local-first routines is the right call for production. you want your CLAUDE.md and architecture rules enforced without sending codebase context to external infra. the multi-agent support is key too — CC for architecture, Codex for review, OpenCode for quick tasks. composable agent stacks running on your own terms.

And governance already available!

Are you joking? That’s an old project, scammer

Open source doesn’t wait for permission. It sees a feature and starts shipping.

damn, that was fast. local routines with swappable agents is HUGE for testing agent behavior, always a pain to debug these things remotely.

the multi-agent composability angle here is underrated. once you start mixing CC for architecture + OpenCode for quick iterations, you really want shared memory across sessions or context bleeds everywhere does exactly this, persists memory locally across all your agents. just updated with native OpenCode + Codes integration if anyone wants to try it (totally free, runs local)

insane!!

seeing OpenClaw support right out of the gate is massive

Insane speed, mad props to the dev, Multica Autopilot that lets you run routines locally with any agent is a total game changer.

Nowadays, whenever Claude releases a new feature, an open-source alternative typically emerges on that very same day and often claiming compatibility with any model.

This is the part that excites me. I've been running exactly this kind of automation locally — scheduled Playwright screenshots of a trading dashboard, API data pulls, auto-composed tweets — all duct-taped together with Task Scheduler and Python scripts. Having a clean framework around it, whether cloud or open-source local, turns those brittle scripts into something you can actually trust to run every morning.

The open source velocity around Claude Code tooling is unreal. Routines barely dropped and people are already building production workflows. The extensibility was clearly designed for exactly this.

That’s so cool! Glad that open source is growing actively, instead of everything being closed behind the paywalls

Mad respect to the dev, local routines in hours is wild, can’t wait to spin up Multica Autopilot and tinker with it.

Open-source velocity's unreal. Local execution with any agent though, that's not a clone. Different value prop entirely.

Running routines locally trades cloud scalability for data privacy, but what's the latency impact on scheduled tasks.

Running this locally changes the game. Avoiding proxy latency and API fees while keeping full control over the agent stack is exactly what the open-source community needed. I’m definitely going to benchmark this against my current local setup.

what a time to be a dev.

Multica Autopilot is slick, but can it beat Claude's built-in speed?