<- all tokdocs

Claude Is Cloning OpenClaw: Remote Control, Scheduled Tasks, Auto Memory

Watch on TikTok

View on TikTok ->

In the last few days Anthropic shipped three Claude features that look suspiciously like OpenClaw in shape: remote control of Claude Code from mobile, scheduled tasks (heartbeat-style wake-ups), and auto memory. The read is clear — Anthropic is absorbing OpenClaw's novel surface area into the first-party Claude experience.

The Three New Features

1. Remote Control — Claude Code from Your Phone

You can now drive a local Claude Code instance from the Claude mobile app. Send instructions from your phone, the agent executes on your computer.

OpenClaw's equivalent: Remote control via Telegram — you DM the bot, the agent runs locally. Anthropic just replaced the Telegram bridge with a first-party mobile app.

2. Scheduled Tasks — Cron for Agents

Claude now supports scheduled task execution. Two use-cases immediately light up:

  • Cron drops — "every Monday at 9am, summarize this repo's new issues"
  • Heartbeat-style wake-ups — agents that self-activate on a schedule and decide what to do next

OpenClaw's equivalent: The "heartbeat" feature where your agent wakes up on an interval and takes action. Anthropic shipped the underlying primitive.

3. Auto Memory — Persistent Context Without Manual Pinning

Claude Code now saves and recalls memory automatically, rather than requiring the user to manually pin context.

OpenClaw's equivalent: OpenClaw's built-in persistent memory. Notably, @promptwarrior says Anthropic's version is "probably a bit better."

The Pattern

OpenClaw primitive Claude first-party equivalent
Telegram remote control Mobile app → local Claude Code
Heartbeat / scheduled wake-ups Scheduled tasks in Claude
Persistent memory Auto memory in Claude Code

This is textbook platform behavior: a hot community tool ships a workflow, the platform-maker builds it into the base product. It's neither new nor surprising — but it matters because OpenClaw's entire value proposition was being the connective tissue Anthropic hadn't shipped yet.

Why Anthropic-Native Wins Over Time

@promptwarrior's prediction — that the Anthropic-native version will beat OpenClaw on security and integration — is the right bet for most users:

  • Security — a first-party feature inherits Anthropic's identity, auth, and data-handling model. No third-party bot token sitting in Telegram with access to your Gmail.
  • Integration — Anthropic can wire scheduled tasks, memory, and remote control into the same context graph Claude Code already uses, versus OpenClaw having to bridge between its own state and Claude's.
  • Latency and reliability — no extra network hop through a third-party bot.

That said, OpenClaw will stay alive for power users who want:

  • Cross-model agent orchestration (not just Claude)
  • Telegram-native UX specifically
  • Self-hosted control over the whole loop

Key Takeaways

  • Anthropic shipped three features in days that mirror OpenClaw's core primitives: remote control, scheduled tasks, auto memory
  • The integration quality of the Anthropic-native version will likely exceed OpenClaw's third-party bridge
  • Expect platforms to routinely absorb community tooling once it validates a use case — OpenClaw was the proof-of-concept, Anthropic is shipping the finished product
  • Watch for Anthropic's official "OpenClaw equivalent" and test it against OpenClaw on security and service integration

Resources

  • Claude Code — Anthropic's terminal-based coding agent
  • Claude Code Docs — Full feature reference including memory and scheduling
  • OpenClaw — The community agent framework being mirrored

Published April 16, 2026. Writeup generated from a favorited TikTok.