#Older version

1 messages · Page 1 of 1 (latest)

keen apex
#

How do I uninstall the new version and install an old version because after updating it to newer version its not working getting stuck in terminal and all

hazy rockBOT
#

To help others find answers, you can mark your question as solved via Right click solution message -> Apps -> âś… Mark Solution

slow canyonBOT
#

First, please run this and paste it so we know what you’re on:

openclaw --version

If you just want to roll back without deleting your config/workspace, do not uninstall ~/.openclaw. Try:

openclaw update --tag 2026.4.23 --yes
openclaw doctor
openclaw gateway restart

If openclaw update itself is stuck/broken, reinstall that exact version with pnpm:

keen apex
#

And there was an error with channel telegram some 5000ms

slow canyonBOT
#

That sounds like Telegram startup/probe timing out, not necessarily the whole install being broken. Please don’t paste your Telegram bot token, but do paste the exact error line if you can.

Run these first:

openclaw --version
openclaw status --all
openclaw channels status --probe

If it’s hanging because Telegram can’t reach the Bot API, check network:

curl -I https://api.telegram.org
openclaw logs --follow
keen apex
#

Im using open router with z-ai free model ⚠️ API rate limit reached. Please try again later. This error can so should I switch to another openrouter model?

slow canyonBOT
#

Yes — that error is from the model/API quota, not usually an OpenClaw install problem. So uninstalling/rolling back probably won’t fix a rate limit.

Try this in the chat with your bot:

/model list

Then pick another configured model:

/model 2

or, if you know the exact OpenRouter model id:

# 
/model openrouter/<model-id>

Free OpenRouter models often have tighter limits, so the practical options are:

  1. Wait until the rate limit resets.
  2. Switch to another OpenRouter model with /model.
  3. Use a paid OpenRouter key/credits if you need reliability.
  4. Ask in #models for current model recommendations — model availability/rate limits change too fast for me to safely recommend a specific one here.

Docs: https://docs.openclaw.ai/concepts/models

keen apex
#

So I need to do openclaw config? And open router api and select a new model

slow canyonBOT
#

Yep — roughly:

  1. Keep the OpenRouter API key configured
    If it’s already set, no need to redo it. If not:

    openclaw onboard --auth-choice openrouter-api-key

  2. Pick/switch model
    In chat with your bot:

    /model list

    Then choose one by number, or set one from CLI:

# 
   openclaw models set openrouter/<provider>/<model>

  1. For free OpenRouter models specifically, you can scan candidates:

    openclaw models scan --set-default

If you’re using free models, rate limits are normal. Switching can help, but a paid OpenRouter key/credits will be more reliable.

Docs: https://docs.openclaw.ai/providers/openrouter

keen apex
#

I have selected nvidia nemotron nano 9b v2 after selecting what do I do

slow canyonBOT
#

If you selected it from /model, you’re basically done.

Next steps:

/model status

Then send a simple test message like:

hi, reply with one sentence

If it still uses the old/rate-limited model, start a fresh session:

# 
/new

If you selected it from terminal instead, verify:

openclaw models status

You only need openclaw config if the OpenRouter API key/model wasn’t saved. If you get rate-limited again, it’s the free OpenRouter model quota — wait or switch to another model with /model list.

keen apex
#

I mean like how do I check it or start it from telegram as I selected a new model

slow canyonBOT
#

Ah got it. Two important bits:

  1. If you only selected the model on OpenRouter’s website, OpenClaw does not automatically know. You must select it in OpenClaw too.
  2. If you selected it in Discord with /model, that may only affect that Discord session. Telegram may need its own /model or a fresh default.

Do this:

openclaw gateway status
openclaw channels status --probe

Then open Telegram, DM your bot:

/start
# 
/model status

If it says pairing required, run on the machine:

openclaw pairing list telegram
openclaw pairing approve telegram <CODE>

Then in Telegram:

/model list
keen apex
#

No i selected it from open claw config again so nect what do to chat in telegram

slow canyonBOT
#

Gotcha — if you selected the model in OpenClaw config, next do this:

openclaw gateway restart
openclaw channels status --probe

Then open Telegram and DM your bot:

/start

If it replies, send:

# 
/model status

Then start a clean chat/session with the new model:

/new

Now just message it normally.

If Telegram says pairing required, run:

openclaw pairing list telegram
keen apex
#

It still says api limit reached should I use a new api from openrouter?

slow canyonBOT
#

Maybe — but a new OpenRouter API key only helps if the limit is on that key.

Usually:

  • Same OpenRouter account + new key → probably still rate-limited if it’s account/free-quota based.
  • Add credits / paid usage → most reliable fix.
  • Switch to another model → helps if the specific free model is rate-limited.
  • Different OpenRouter account/key → may work, but don’t abuse free limits.

If you do get a new key, configure it again:

openclaw onboard --auth-choice openrouter-api-key
openclaw gateway restart

Then in Telegram:

# 
/new

Also check your OpenRouter dashboard usage/limits. And don’t paste the API key here.

keen apex
#

Now it says llm request timeout

slow canyonBOT
#

That’s a provider/model timeout now — not an uninstall problem.

Try this:

openclaw models status --probe

If OpenRouter is timing out, either switch model again or give it more time:

openclaw config set models.providers.openrouter.timeoutSeconds 120 --strict-json
openclaw gateway restart

Then in Telegram:

# 
/new

and test with a tiny message like “hi”.

If it still times out, the free Nemotron route is probably overloaded/slow on OpenRouter. Use /model list and pick another model, or ask in #models for current good OpenRouter choices. Also please paste openclaw --version if it keeps happening. Docs: https://docs.openclaw.ai/concepts/models

keen apex
#

Should I switch account and try once?

slow canyonBOT
#

You can try once, but I’d do it in this order:

  1. Try a different OpenRouter model first — timeouts often mean that specific free route is overloaded.
  2. Check OpenRouter dashboard for quota/rate-limit state.
  3. If you have another legitimate OpenRouter account/key, you can test it, but don’t rely on account switching to dodge free limits. Paid credits are the stable fix.

To switch key:

openclaw onboard --auth-choice openrouter-api-key
openclaw gateway restart

Then in Telegram:

# 
/new

If it still says timeout, it’s probably the model/provider route, not the account.

keen apex
#

I have used new api and selected minimax 2.5 next what to do

slow canyonBOT
#

Nice — now just restart and test from Telegram:

openclaw gateway restart
openclaw channels status --probe
openclaw models status --probe

Then in Telegram DM with your bot:

/new
hi, reply in one sentence

If it replies, you’re good.

#

If it still fails, paste the exact error text from:

openclaw logs --follow

Don’t paste API keys/tokens.

keen apex
#

Its not responding at all?

slow canyonBOT
#

Okay, now we need to see where it’s failing: gateway, Telegram, or model.

Run these and paste the output:

openclaw --version
openclaw status
openclaw gateway status
openclaw channels status --probe
openclaw models status --probe

Then restart once:

openclaw gateway restart