No Response from Bot | Friends of the Crustacean 🦞🤝 | Page 1

opaque quartz Apr 20, 2026, 9:02 PM

#

Hi, my bot is not responding to my messages. I’ve already reinstalled OpenClaw three times and configured OpenRouter, but it’s still not working. I also tried different models, but nothing happens. Two days ago, OpenClaw was still working fine. Could you please help me set it up?

sturdy islandBOT Apr 20, 2026, 9:02 PM

#

junior hillBOT Apr 20, 2026, 9:02 PM

#

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

outer sinewBOT Apr 20, 2026, 9:03 PM

#

That’s frustrating. Don’t reinstall again yet, this is usually a status/logs issue, not a reinstall issue.

Please paste these outputs in one reply:

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

Then, in another terminal:

#

openclaw logs --follow

While logs are running, send the bot one test message, then paste the first errors or warnings you see.

Also tell me:

which channel you’re using, Discord / Telegram / WhatsApp / etc.
if this is DM or a group/server channel
whether you updated OpenClaw around the time it stopped working

On Discord specifically, one very common cause is mention gating, logs often show:

drop guild message (mention required)
pairing / allowlist blocks
channel transport disconnected

Docs I’m following:

#

If you installed from source, use pnpm openclaw ... instead of openclaw .... Once you paste the outputs, I’ll help narrow it down fast.

opaque quartz Apr 20, 2026, 9:25 PM

#

gateway connect failed: GatewayClientRequestError: pairing required
21:20:32+00:00 warn gateway {"subsystem":"gateway"} security audit: device access upgrade requested reason=scope-upgrade device=0631289dc5953d7d6536c022763ae00ce4075f95d9143a71a34a0c60d9b9e380 ip=unknown-ip auth=token roleFrom=operator roleTo=operator scopesFrom=operator.read scopesTo=operator.admin,operator.approvals,operator.pairing,operator.read,operator.talk.secrets,operator.write client=cli conn=def54ad2-cdfb-4489-8f28-0e7e64e7867f
21:20:34+00:00 error gateway connect failed: GatewayClientRequestError: pairing required
21:20:34+00:00 error gateway connect failed: GatewayClientRequestError: pairing required

#

Discord
2.Server
No

outer sinewBOT Apr 20, 2026, 9:26 PM

#

Nice, this log is actually very helpful. We found the issue.

This is not an OpenRouter/model problem. Your client is hitting a device scope upgrade and the gateway is refusing it until you approve the new access.

pairing required here means gateway device pairing, not chat sender pairing.

Run:

openclaw devices list
openclaw devices approve <requestId>

Important: openclaw devices approve --latest only previews. You need the exact requestId from devices list.

Your log shows the CLI currently has only:

#

operator.read

and it is now requesting:

operator.admin
operator.approvals
operator.pairing
operator.read
operator.talk.secrets
operator.write

So approve that pending upgrade, then retry your command.

After that, send me:

#

openclaw --version
openclaw status
openclaw gateway status

If you’re running from source, use pnpm openclaw ... instead.

Docs:

If devices approve succeeds and you still get pairing required, paste the new openclaw devices list output and we’ll go one step deeper.

#

Perfect, that narrows it down.

Because this is Discord in a server, I’d check two things:

Fix the gateway pairing/scope-upgrade first

openclaw devices list
openclaw devices approve <requestId>

Then test Discord server routing
- send a message that explicitly @mentions the bot
- keep logs open while testing:
```
openclaw --version
openclaw channels status --probe
openclaw logs --follow
```

#


Please paste those outputs after you test.

For Discord server replies, the most common causes are:
- **Message Content Intent** not enabled
- guild/channel not allowed
- `requireMention` is on, so non-mentions are ignored

The docs match that exactly:
- <https://docs.openclaw.ai/channels/discord#troubleshooting>
- <https://docs.openclaw.ai/channels/troubleshooting>
- <https://docs.openclaw.ai/gateway/troubleshooting#no-replies>

So the immediate next step is:

#

openclaw devices list
openclaw devices approve <requestId>
openclaw channels status --probe
openclaw logs --follow

Then mention the bot once in the server and paste whatever the logs say. If it shows mention required, blocked, or an intent error, I can tell you the exact fix.

opaque quartz Apr 20, 2026, 9:30 PM

#

Direct scope access failed; using local fallback.
Pending (1)
┌──────────────────────────────────────┬────────────────────────────────────┬──────────┬──────────────────────────────────────────────────────────────────────────────┬────────────┬──────────┬────────┐
│ Request │ Device │ Role │ Scopes │ IP │ Age │ Flags │
├──────────────────────────────────────┼────────────────────────────────────┼──────────┼──────────────────────────────────────────────────────────────────────────────┼────────────┼──────────┼────────┤
│ bba9a6a0-9ae3-4ace-a7c8-3f8efddf0a7d │ 0631289dc5953d7d6536c022763ae00ce4 │ operator │ operator.admin, operator.read, operator.write, operator.approvals, operator. │ │ just now │ repair │
│ │ 075f95d9143a71a34a0c60d9b9e380 │ │ pairing, operator.talk.secrets │ │ │ │
└──────────────────────────────────────┴────────────────────────────────────┴──────────┴──────────────────────────────────────────────────────────────────────────────┴────────────┴──────────┴────────┘
Paired (2)

│ Device │ Roles │ Scopes │ Tokens │ IP │
────────────────────────────────────┼─────────────┼────────────┤
│ gateway:doctor.memory.status │ operator │ operator.read │ operator │ │
│ 5b9256f0b58701e8792d3b827723da9fd29b66e6d777be76856e0f5e69bef291 │ operator │ operator.admin, operator.read, operator.write, operator.approvals, operator.pairing │ operator │ │

outer sinewBOT Apr 20, 2026, 9:31 PM

#

Yep, that’s the pending upgrade.

Approve this exact request:

openclaw devices approve bba9a6a0-9ae3-4ace-a7c8-3f8efddf0a7d

Then run:

openclaw --version
openclaw channels status --probe
openclaw logs --follow

After logs --follow is running, go into the Discord server and @mention the bot once, then paste the new log lines.

#

Direct scope access failed; using local fallback. is not the main problem here. The real blocker is the pending device scope repair/upgrade.

If the approve command works but the bot is still silent in the server, the next likely Discord-side causes are:

requireMention
guild/channel allowlist
Message Content Intent

Docs:

Paste the result of the approve command plus the next log lines, and I’ll help with the next step.

opaque quartz Apr 20, 2026, 9:33 PM

#

21:31:30+00:00 info Approved 0631289dc5953d7d6536c022763ae00ce4075f95d9143a71a34a0c60d9b9e380 (bba9a6a0-9ae3-4ace-a7c8-3f8efddf0a7d)
21:32:42+00:00 info gateway/ws {"subsystem":"gateway/ws"} ⇄ res ✓ channels.status 615ms conn=320e496b…5fcb id=ade4c1e1…df6c
21:32:42+00:00 info Gateway reachable.

Discord default: enabled, configured, running, connected, in:12m ago, bot:@Izzy, token:*** intents:content=limited, works

Tip: status --deep adds gateway health probes to status output (requires a reachable gateway).
Log tail truncated (increase --max-bytes).
21:32:52+00:00 warn agent/embedded {"subsystem":"agent/embedded"} incomplete turn detected: runId=b319b533-702d-4df5-8179-68c52949e84b sessionId=c9a4e54e-33bc-4b6e-8b81-366dba473956 stopReason=stop payloads=0 — surfacing error to user

outer sinewBOT Apr 20, 2026, 9:35 PM

#

Nice, progress. Discord itself is now connected and working.

What the logs say now:

Channel side is healthy
- Discord ... connected ... works
Your bot has limited server visibility
- intents:content=limited
- On Discord, that usually means Message Content Intent is not fully enabled, so in servers the bot may only reliably see @mentions / DMs, not normal messages.
The agent is starting, but the model run is failing
- incomplete turn detected ... payloads=0
- That means OpenClaw received the message, started the run, and then the model returned no visible answer.

So there are 2 remaining fixes:

#

A. For server messages, enable Message Content Intent

In the Discord Developer Portal for your bot, turn on:

Message Content Intent
restart the gateway afterward

Docs:

https://docs.openclaw.ai/channels/discord

B. Test the model directly

Please run and paste:

openclaw --version
openclaw models status
openclaw config get agents.defaults.model.primary

#

openclaw infer model run --model <paste-the-model-id-here> --prompt "Reply with exactly: hello" --json

If that direct infer model run also fails or comes back empty, then the problem is the selected model/provider path, not Discord.

For now, while content=limited is still showing, test in the server by @mentioning the bot. Without full Message Content Intent, plain server messages can be invisible.

Relevant docs:

Paste the models status and infer model run output next, and we can pin down the last blocker.

opaque quartz Apr 20, 2026, 9:39 PM

#

Config : ~/.openclaw/openclaw.json
Agent dir : ~/.openclaw/agents/main/agent
Default : openrouter/arcee-ai/trinity-large-preview:free
Fallbacks (0) : -
Image model : -
Image fallbacks (0): -
Aliases (1) : OpenRouter -> openrouter/auto
Configured models (2): openrouter/auto, openrouter/arcee-ai/trinity-large-preview:free

Auth overview
Auth store : ~/.openclaw/agents/main/agent/auth-profiles.json
Shell env : off
Providers w/ OAuth/tokens (0): -

openrouter effective=profiles:~/.openclaw/agents/main/agent/auth-profiles.json | profiles=1 (oauth=0, token=0, api_key=1) | openrouter:default=sk-or-v1...3d3319d2 | env=sk-or-v1...3d3319d2 | source=env: OPENROUTER_API_KEY
perplexity effective=env:sk-or-v1...3d3319d2 | env=sk-or-v1...3d3319d2 | source=env: OPENROUTER_API_KEY

OAuth/token status

none

outer sinewBOT Apr 20, 2026, 9:40 PM

#

This helps a lot.

Your auth looks fine. So the remaining blockers are now pretty clear:

Discord is still running with limited content intent
- intents:content=limited
- In a server, that usually means the bot will not reliably see normal messages unless it’s @mentioned
- If you want it to answer plain server messages, enable Message Content Intent in the Discord Developer Portal, then restart the gateway
Your current default model is the next likely problem
- openrouter/arcee-ai/trinity-large-preview:free
- Your earlier log, incomplete turn detected ... payloads=0, means the agent started but the model returned no visible answer

So I’d test the model path directly now:

#

openclaw infer model run --model openrouter/arcee-ai/trinity-large-preview:free --prompt "Reply with exactly: hello" --json
openclaw infer model run --model openrouter/auto --prompt "Reply with exactly: hello" --json

Please paste both outputs.

If the first one is empty/fails and openrouter/auto works, then the issue is basically the free preview model, not your Discord setup.

For the Discord side, docs are here:

For model behavior:

#

So at this point, I would:

enable Message Content Intent
test both infer commands above
keep using @mentions in the server until the intent is fixed

Paste the two infer model run results and we can pin it down completely.

opaque quartz Apr 20, 2026, 9:42 PM

#

23:42:18 [agent/embedded] incomplete turn detected: runId=8787c450-9aa8-43c2-b942-d6bf2cc17f2a sessionId=8787c450-9aa8-43c2-b942-d6bf2cc17f2a stopReason=stop payloads=0 — surfacing error to user
{
"ok": true,
"capability": "model.run",
"transport": "local",
"provider": "openrouter",
"model": "arcee-ai/trinity-large-preview:free",
"attempts": [],
"outputs": [
{
"text": "⚠️ Agent couldn't generate a response. Please try again.",
"mediaUrl": null
}
]
}

outer sinewBOT Apr 20, 2026, 9:43 PM

#

Good, that confirms the root cause.

Your current model is the problem:

openrouter/arcee-ai/trinity-large-preview:free
even a direct infer model run returns:
- ⚠️ Agent couldn't generate a response. Please try again.

So at this point, Discord is not the main blocker anymore. The bot is connected, but that model is not producing a usable reply.

Do this next

Switch off that model:

openclaw models set openrouter/auto

#

Then test it directly:

openclaw infer model run --model openrouter/auto --prompt "Reply with exactly: hello" --json

If that works, test the bot again in Discord with an @mention.

One more Discord fix

Your status also showed:

intents:content=limited

So if you want the bot to respond to normal server messages, also enable Message Content Intent in the Discord Developer Portal. Until then, test with @mentions.

Short version

You had two separate issues:

#

gateway/device pairing scope issue, now fixed
bad/default model returning empty replies, this is the current blocker

Docs:

If openrouter/auto also fails, paste that output next. If it works, you’re basically unblocked.

opaque quartz Apr 20, 2026, 9:47 PM

#

23:46:42 [agent/embedded] incomplete turn detected: runId=8787c450-9aa8-43c2-b942-d6bf2cc17f2a sessionId=8787c450-9aa8-43c2-b942-d6bf2cc17f2a stopReason=stop payloads=0 — surfacing error to user
{
"ok": true,
"capability": "model.run",
"transport": "local",
"provider": "openrouter",
"model": "openrouter/auto",
"attempts": [],
"outputs": [
{
"text": "⚠️ Agent couldn't generate a response. Please try again.",
"mediaUrl": null
}
]
}

outer sinewBOT Apr 20, 2026, 9:48 PM

#

That confirms it, this is no longer a Discord issue.

Because both of these fail on a direct model run:

openrouter/arcee-ai/trinity-large-preview:free
openrouter/auto

the break is now isolated to the OpenRouter provider path itself.

Next best check

Please run:

openclaw --version
openclaw models status --probe --probe-provider openrouter

#

Then, in another terminal:

openclaw logs --follow

While logs are open, run this once more:

openclaw infer model run --model openrouter/auto --prompt "Reply with exactly: hello" --json

and paste the OpenRouter-related log lines.

Why this matters

models status --probe should tell us if OpenClaw sees this as:

#

auth
rate_limit
billing
timeout
format
unknown

Right now all we know is: OpenClaw can reach the provider path, but the provider/model run ends with an empty/incomplete response.

Separate Discord note

You still also want to fix:

intents:content=limited

because in server channels that means the bot may only reliably work on @mentions until Message Content Intent is enabled.

Docs:

#

If the probe comes back weird and you’re not on the latest version, I’d update next. Paste the probe output and version first, though.

opaque quartz Apr 20, 2026, 9:53 PM

#

21:52:34+00:00 warn agent/embedded {"subsystem":"agent/embedded"} incomplete turn detected: runId=probe-fafaf0ce-d58b-4f0a-9848-43c37500b18b sessionId=probe-openrouter-9b0d86e6-9714-4be1-93e4-ca30927d3de8 stopReason=stop payloads=0 — surfacing error to user
21:53:39+00:00 warn agent/embedded {"subsystem":"agent/embedded"} incomplete turn detected: runId=8787c450-9aa8-43c2-b942-d6bf2cc17f2a sessionId=8787c450-9aa8-43c2-b942-d6bf2cc17f2a stopReason=stop payloads=0 — surfacing error to user

outer sinewBOT Apr 20, 2026, 9:54 PM

#

We’re close, but I still need the actual probe output, not just the log tail.

Those log lines only confirm the same symptom again: OpenRouter is returning an empty/incomplete response.

Please paste the direct output of:

openclaw --version
openclaw models status --probe --probe-provider openrouter --plain

If --plain is not supported on your version, just run:

openclaw models status --probe --probe-provider openrouter

#

What I need to see is whether OpenClaw classifies OpenRouter as:

ok
auth
rate_limit
billing
timeout
format
unknown

Right now, the strongest read is:

Discord transport is fine
pairing is fixed
your OpenRouter model path is failing even on direct probe/model-run

So the remaining issue is likely either:

#

an OpenRouter-side/provider issue, or
a version-specific OpenClaw bug

Please paste those two command outputs next. If the probe still comes back weird and your version is old, I’d update immediately. If you’re already current, the next step will probably be a GitHub issue with the probe output attached:
https://github.com/openclaw/openclaw/issues

opaque quartz Apr 20, 2026, 9:56 PM

#

🦞 OpenClaw 2026.4.15 (041266a) — I'll butter your workflow like a lobster roll: messy, delicious, effective.

│
◇
Config : ~/.openclaw/openclaw.json
Agent dir : ~/.openclaw/agents/main/agent
Default : openrouter/openrouter/auto (from openrouter/auto)
Fallbacks (0) : -
Image model : -
Image fallbacks (0): -
Aliases (1) : OpenRouter -> openrouter/auto
Configured models (2): openrouter/auto, openrouter/arcee-ai/trinity-large-preview:free

Auth overview
Auth store : ~/.openclaw/agents/main/agent/auth-profiles.json
Shell env : off
Providers w/ OAuth/tokens (0): -

openrouter effective=profiles:~/.openclaw/agents/main/agent/auth-profiles.json | profiles=1 (oauth=0, token=0, api_key=1) | openrouter:default=sk-or-v1...3d3319d2 | env=sk-or-v1...3d3319d2 | source=env: OPENROUTER_API_KEY
perplexity effective=env:sk-or-v1...3d3319d2 | env=sk-or-v1...3d3319d2 | source=env: OPENROUTER_API_KEY

OAuth/token status

none

Auth probes
┌────────────────────────────┬──────────────────────────────┬────────────┐
│ Model │ Profile │ Status │
├────────────────────────────┼──────────────────────────────┼────────────┤
│ openrouter/openrouter/auto │ openrouter:default (api_key) │ ok · 21.8s │
└────────────────────────────┴──────────────────────────────┴────────────┘
Probed 1 target in 41.7s

outer sinewBOT Apr 20, 2026, 9:58 PM

#

Okay, we’ve got it.

This looks like a known OpenRouter bug in OpenClaw, not a bad setup on your side.

Why I’m confident:

Version: 2026.4.15
Auth probe: ok
Discord transport: connected / works
Pairing: fixed
Direct model run still fails with payloads=0
That matches existing reports almost exactly

Relevant issues:

#

So the current picture is:

your OpenRouter key is valid
OpenClaw can reach OpenRouter
OpenClaw is dropping/losing the returned assistant payload on this provider path

So no, I would not keep reinstalling. I don’t think this is fixable with more local setup changes.

Best workaround right now

If you need it working today, use a non-OpenRouter provider path temporarily in OpenClaw.

What I’d do next

subscribe / comment on one of those issues with your repro
include:
- 2026.4.15
- models status --probe showing openrouter ... ok
- the infer model run output with payloads=0

#

If you want, I can help you switch to a temporary non-OpenRouter provider next.

Small separate note: once the model issue is out of the way, you should still enable Message Content Intent for normal Discord server messages, otherwise server replies may only work reliably on @mentions.

opaque quartz Apr 20, 2026, 9:59 PM

#

whatist the best way to install openclaw on linux?

outer sinewBOT Apr 20, 2026, 10:00 PM

#

On Linux, the best default install is:

curl -fsSL https://openclaw.ai/install.sh | bash
openclaw onboard --install-daemon

If you want the most debuggable / support-friendly setup, I’d actually use the git install instead:

curl -fsSL https://openclaw.ai/install.sh | bash -s -- --install-method git
openclaw onboard --install-daemon

My recommendation:

just want it working: normal installer
want easier debugging / local fixes / better support: git install

#

A couple of Linux notes from the docs:

use Node 22+
pnpm is recommended
Bun is not recommended for the gateway

Docs:

Also, for the issue we just traced in this thread, reinstalling won’t fix the current OpenRouter bug on 2026.4.15. That part is separate.

opaque quartz Apr 20, 2026, 10:01 PM

#

i just need to run this

curl -fsSL https://openclaw.ai/install.sh | bash
openclaw onboard --install-daemon

outer sinewBOT Apr 20, 2026, 10:01 PM

#

Yep, for the standard Linux install, that’s the right path:

curl -fsSL https://openclaw.ai/install.sh | bash
openclaw onboard --install-daemon

A couple quick notes:

run it on the Linux machine that will host the gateway
onboarding will walk you through auth/channels
for Discord server replies, also enable Message Content Intent

Docs:

Just one caveat: for the specific OpenRouter bug we found above, reinstalling with those commands probably won’t fix that part.

opaque quartz Apr 20, 2026, 10:03 PM

#

CLI still installed. Remove via npm/pnpm if desired.

outer sinewBOT Apr 20, 2026, 10:03 PM

#

That message is fine.

It just means you already have the OpenClaw CLI installed globally. You do not need to remove it to continue.

You can just keep going with:

openclaw onboard --install-daemon

Only remove the old CLI if you specifically want a totally clean reinstall. Otherwise, that line is just informational.

opaque quartz Apr 20, 2026, 10:04 PM

#

i want to total clean reinstall

outer sinewBOT Apr 20, 2026, 10:05 PM

#

If you want a true clean reinstall, I’d use the built-in commands, not manual file deletion.

Cleanest full wipe

This removes the service, config, state, and CLI-managed install:

openclaw uninstall --all --yes --non-interactive
curl -fsSL https://openclaw.ai/install.sh | bash
openclaw onboard --install-daemon

Slightly lighter option

If you want to wipe OpenClaw’s state but keep the CLI installed:

openclaw reset --scope full --yes --non-interactive

#

openclaw onboard --install-daemon

If you care about your current config/credentials, back up ~/.openclaw first.

Docs:

One important caveat: based on everything we traced above, a clean reinstall still probably won’t fix the current OpenRouter bug on 2026.4.15.

opaque quartz Apr 20, 2026, 10:13 PM

#

how to configure openrouter by setup openclaw

outer sinewBOT Apr 20, 2026, 10:14 PM

#

The setup path for OpenRouter in OpenClaw is:

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

Then:

choose OpenRouter
paste your API key from https://openrouter.ai/keys
finish onboarding

By default, onboarding uses:

openrouter/auto

#

If you want to switch later to a specific OpenRouter model:

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

Example format only:

openrouter/google/gemini-2.5-flash

To verify config afterward:

openclaw models status

#

Docs:

One important note for your current case: even if you set it up correctly, the OpenRouter bug we traced above on 2026.4.15 can still cause payloads=0. So setup and runtime are two separate things here.

opaque quartz Apr 20, 2026, 11:08 PM

#

The model did not produce a response before the LLM idle timeout. Please try again, or increase agents.defaults.llm.idleTimeoutSeconds in your config (set to 0 to disable).

outer sinewBOT Apr 20, 2026, 11:09 PM

#

That message means the model sat idle long enough that OpenClaw gave up waiting.

The config key is real, docs here:
https://docs.openclaw.ai/concepts/agent-loop

If you want to test a longer wait, try:

openclaw config set agents.defaults.llm.idleTimeoutSeconds 300

If you want to disable that watchdog completely:

openclaw config set agents.defaults.llm.idleTimeoutSeconds 0

#

I’d try 300 first, not 0.

That said, in your thread I think this is still probably another symptom of the same OpenRouter bug we already isolated, not just a normal timeout. So increasing the timeout may help testing, but I would not expect it to fully fix the payloads=0 issue on 2026.4.15.

opaque quartz Apr 21, 2026, 7:36 AM

#

The issue appears to be the OpenRouter base URL in:
~/.openclaw/agents/main/agent/models.json

It was set to:
https://openrouter.ai/v1

Changing it to:
https://openrouter.ai/api/v1

resolved the problem.

#No Response from Bot

A. For server messages, enable Message Content Intent

B. Test the model directly

Do this next

One more Discord fix

Short version

Next best check

Why this matters

Separate Discord note

Best workaround right now

What I’d do next

Cleanest full wipe

Slightly lighter option