Blog / troubleshooting
Why Hermes Agent Is Not Sending or Receiving Messages
When Hermes Agent is not sending or receiving messages, start with the channel connection before changing prompts or models. Most failures come from credentials, permissions, QR sessions, runtime state, or model API key configuration rather than the assistant logic itself.
- 01
Check the channel
Confirm the Telegram token, Discord permissions, or WhatsApp QR session is still valid.
- 02
Check runtime state
Verify that the Hermes Agent deployment is running and not stuck in a restart or redeploy state.
- 03
Check model access
Confirm the selected model and API key are valid, available, and within usage limits.
- 04
Inspect logs
Use logs and recent test messages to identify whether the failure is inbound, outbound, or model-related.
Four-step path
How do you troubleshoot message failures?
Work from the outside in: channel, runtime, model access, logs, then prompt behavior.
- Step 01
Check the channel
Confirm the Telegram token, Discord permissions, or WhatsApp QR session is still valid.
- Step 02
Check runtime state
Verify that the Hermes Agent deployment is running and not stuck in a restart or redeploy state.
- Step 03
Check model access
Confirm the selected model and API key are valid, available, and within usage limits.
- Step 04
Inspect logs
Use logs and recent test messages to identify whether the failure is inbound, outbound, or model-related.
What should you check first when Hermes Agent does not reply?
Check the channel connection first, because Hermes Agent cannot reply if it cannot receive the message or post back into the conversation.
A broken token, missing Discord channel access, or disconnected WhatsApp session can make the assistant look offline even when the runtime is working.
How do you tell inbound and outbound failures apart?
Inbound failures mean Hermes Agent never receives the message; outbound failures mean it receives the event but cannot send the reply back to the channel.
- If no event appears in logs, inspect credentials and channel permissions.
- If an event appears but no reply is sent, inspect send permissions and channel state.
- If a model error appears, inspect API keys, quota, and model availability.
- If the runtime restarts repeatedly, inspect deployment health before channel settings.
Can model API keys stop messages from sending?
Yes. A missing, revoked, quota-limited, or provider-specific model API key can stop Hermes Agent after the channel receives the message.
This is why troubleshooting should separate channel delivery from model execution. A message can reach the runtime and still fail if the model request cannot complete.
What is the safest test message?
Use a short low-risk message in a private Telegram chat, private Discord channel, or known WhatsApp test conversation, then compare the timestamp with deployment logs.
Testing with a single controlled message prevents noise from live users and gives a clean timestamp for log inspection.
Which primary sources support this page?
Product behavior is checked against the current deployment flow. These external links provide the upstream project or channel documentation used for setup details.
Direct answers
Frequently asked questions
Should I change the prompt first when replies fail?
No. Check channel access, runtime health, and model API access before changing prompts.
Why does Hermes Agent receive messages but not answer?
The runtime may receive the event but fail on model execution, outbound channel permission, quota, or reply routing.
Why does Hermes Agent work in private chat but not a group?
Group-specific privacy settings, role permissions, or channel overrides are usually involved when private chat works but group chat fails.
Can restarting fix message failures?
Restarting can clear transient runtime issues, but it will not fix a revoked token, missing permission, disconnected QR session, or invalid model key.
Key takeaways
- When Hermes Agent is not sending or receiving messages, start with the channel connection before changing prompts or models. Most failures come from credentials, permissions, QR sessions, runtime state, or model API key configuration rather than the assistant logic itself.
- Start with one model, one channel, and one controlled test conversation.
- Use managed hosting when deployment speed matters more than operating the server yourself.
Continue the setup
Related Hermes Agent guides
Blog guide
How to Set Up a BotFather Token for Hermes Agent
Create a Telegram BotFather token for Hermes Agent, protect the credential, connect it to a managed deployment, and verify the first reply.
Read nextBlog guide
How to Configure Discord Permissions for Hermes Agent
Configure Discord bot permissions, gateway intents, roles, and channel access before deploying Hermes Agent into a server.
Read nextBlog guide
How to Connect Hermes Agent with a WhatsApp QR Session
Connect Hermes Agent to WhatsApp with a QR linked-device session, verify the runtime, and recover cleanly when the session expires.
Read nextBlog guide
How to Restart or Redeploy Hermes Agent Safely
Restart or redeploy Hermes Agent safely by checking runtime state, active sessions, credentials, and message flow before returning to production use.
Read next