Show HN: Script to bulk delete Claude chats from the web UI

Who Should Read

A developer with hundreds of claude.ai chats who finds deleting them one by one tedious

Core Mechanics

• Works around 'Select all' limit
  • The 'Select all' button on /recents only covers rows rendered on screen.
  • The script was created because of that limit.
• Paste into the console
  • Open claude.ai/recents and press F12 for the console.
  • Paste the contents of delete-all.js and hit Enter.
  • Confirm one dialog per organization.
• Deletion takes minutes
  • Conversations disappear from the UI slowly, which is said to be normal.
  • Keep the tab open until the console shows Finished.
• Uses undocumented endpoints
  • These are internal endpoints inferred from browser network requests, not public Anthropic APIs.
  • The README warns they may change or stop working at any time.
• Own account only, unaffiliated
  • It is not affiliated with Anthropic.
  • The README says to use it only with your own Claude account.
  • It is MIT licensed.

Community Reactions

• Wanting to keep a few chats
  • One commenter said ChatGPT is harder and they want to keep a couple of reference chats, so nuke-all doesn't fit.
• Incognito chat as default
  • A commenter uses the temporary chats in Claude and ChatGPT as their default since they vanish when the tab closes.
  • They note you can't save such a chat once started.
• Deleted may not mean gone
  • Commenters pointed out it's unclear that deleting actually removes data from Anthropic's datastore.
  • Some suggested local AI for sensitive content.
• No multi-select delete
  • Commenters complained that AI chat products lack multi-select delete, offering only delete-all or per-chat confirmation.

How to Apply

• Switch to incognito chats by default
  • Start one-off questions in Claude's temporary chat (ghost icon, top right) so no history builds up.
• Back up chats before running it
  • The script deletes everything, so copy any chats you want to keep first.
  • Read the code and check what requests it sends before running.
• Move sensitive work to a local model
  • Run conversations you don't want on external servers locally, e.g. with llama-cli.

Try It

# Paste into Claude or ChatGPT (to safety-check the script)

Below is JavaScript I plan to paste into a browser console. Review it before I run it.
1. List which URLs and HTTP requests it sends.
2. Tell me whether it reads other data or sends anything externally besides deleting.
3. Tell me which actions are irreversible.

{{contents of delete-all.js}}

Terminology

Related Papers

• DeepInfer: High-performance inference framework for DeepSeek models

  DeepSeek open-sourced DeepInfer, its production inference engine, under Apache-2.0.

• Mercury 3: Diffusion at scale

  Scaling diffusion LLM 40x gives Mercury 3: 4–8x faster than speed-tier models, similar quality

• DeepSeek banned from U.S. federal civilian agencies under new DHS order

  DHS order bans DeepSeek from U.S. federal civilian agency IT systems

• Timestral: Mistral's time-series foundation models

  A 209M time-series model ranks first on three benchmarks and runs on CPU.

• DeepSeek V4.1 Flash 1017 (Max)

  DeepSeek V4.1 Flash 1017 Max scores 45 on the Intelligence Index at 211 tok/s and $0.30 input.

• If you don't run your own LLM proxy, how do you know whats going on?

  Running your own LLM proxy shows which keys, models, and cache misses drive your usage.

• DeepSeek: DeepSeek v4.1 Flash 1017

  DeepSeek V4.1 Flash 1017: $0.05/M input, 1M context, built for agents

• DeepSeek-v4.1-Flash-1017

  DeepSeek V4.1 Flash minor update lifts coding and agent benchmarks while cutting run time and output length

Related Resources

• https://github.com/MatteoLeonesi/bulk-delete-claude-chat