>>180879, >>180880, >>180881, >>180882, >>180883, >>180884, >>180885, >>180886, >>180887, >>180888, >>180889, >>180890, >>180891, >>180892, >>180893, >>180894, >>180895, >>180896, >>180897, >>180898, >>180899, >>180900, >>180901, >>180902, >>180903, >>180904, >>180905, >>180906, >>180907, >>180908, >>180909, >>180910, >>180911, >>180912, >>180913, >>180914, >>180915, >>180916, >>180917, >>180918, >>180919, >>180920, >>180921, >>180922, >>180923, >>180924, >>180925, >>180926, >>180927, >>180928, >>180929, >>180930, >>180931, >>180932, >>180933, >>180934, >>180935, >>180936, >>180937, >>180938, >>180939, >>180940, >>180941, >>180942, >>180943, >>180944, >>180945, >>180946, >>180947, >>180948, >>180949, >>180950, >>180951, >>180952, >>180953
Todd Saunders @toddsaunders - Every company in the world has a graveyard of automations...at least we did.
Routines replaces all of that with a single primitive.
A trigger, a context window, and an agent that can actually reason about what it's looking at.
That's a fundamentally different architecture than "if this then that."
The API trigger is also incredibly interesting.
That means any system in your stack can invoke a reasoning agent on demand. Think things like a webhook from Stripe, a failed deploy, a customer cancellation.
Anything that currently sends a Slack notification and hopes a human reads it can now trigger an agent that actually does something about it.
Quote:
Noah Zweben @noahzweben
Video: Claude Code Routines are here! In addition to a schedule, you can now trigger templated agents via GitHub event or API – with our infra & your MCP+repos
They've changed how we do docs, backlog maintenance and more internally at Anthropic
Get started at
https://claude.ai/code/routines
https://x.com/toddsaunders/status/2044238577899295070

Todd Saunders @toddsaunders - I’ve been trying to be friends with Joe for years… never was I able to get him to respond to a text