Troubleshooting

Install fails: EACCES: permission denied, open '/home/node/.n8n/nodes/package.json'

n8n can't write into its own custom-nodes directory. This is a host-side permissions issue, not a package issue.

[!WARNING] Don't chmod 777 to "fix" this — it'll work but anyone on the host can then write into your n8n data. Use chown to the node user instead (UID 1000 by default).

Docker fix (most common):

# Find your n8n container
docker ps | grep n8n

# Fix permissions on the in-container path
docker exec -u 0 <n8n-container> sh -c "mkdir -p /home/node/.n8n/nodes && chown -R node:node /home/node/.n8n"

# If the volume is on the host
sudo chown -R 1000:1000 ~/.n8n
sudo chmod -R u+rwX ~/.n8n

# Restart n8n
docker restart <n8n-container>

Manual install (sidesteps the UI):

docker exec -u node -it <n8n-container> sh
cd ~/.n8n && mkdir -p nodes && cd nodes
echo '{"dependencies":{}}' > package.json
npm install n8n-nodes-tellzm
exit
docker restart <n8n-container>

"Sign in with OAuth2" fails / "Credential is invalid"

• Confirm you're signed into Tellzm in the same browser. The OAuth flow redirects you to tellzm.com/oauth/authorize — if you're not signed in there, you'll be bounced to sign-in first.
• If your n8n is behind a tunnel or non-standard reverse-proxy, ensure the redirect URL that n8n displays matches https://<your-n8n-host>/rest/oauth2-credential/callback. Otherwise the OAuth server will refuse to redirect back.

"This token is bound to workspace X, not Y"

You're hitting the multi-workspace policy guard. The credential you used wasn't authorized for the workspace you selected on the node. Two fixes:

• On the node: pick a different workspace from the Workspace dropdown (it shows only workspaces this credential is authorized for).
• Re-authorize: in Tellzm → Settings → AI Assistants → revoke the credential, then re-create it in n8n and include the missing workspace in the consent screen.

Trigger node never fires

• Confirm the trigger is active in n8n (workflow toggle is on).
• Click Listen for test event while inactive — does Tellzm forward the event? If not, the webhook subscription wasn't created. Disable + re-enable the node to re-register.
• If you have a filter set, temporarily clear it. The filter runs server-side before the webhook is sent.
• Check the rate limit: each credential is capped at 600 requests/minute. Heavy fan-out workflows can hit this — the server returns 429 Retry-After and the package retries once. Successive throttles fail the run.

Rate-limit errors (429)

The Tellzm side enforces 600 requests/min per credential. The package automatically retries once after Retry-After. If you see "Axon API still rate-limited after Retry-After (Ns)" in n8n, the workflow is sustained over the cap — split it into smaller batches or use the Trigger node instead of polling.

"Workspace not found" after creating a new workspace

If you enabled auto-include future workspaces during consent, new workspaces should be reachable automatically. If you did NOT enable that, you need to re-authorize the credential and tick the new workspace.

A quick check: hit GET /mcp/whoami via the Raw API operation and look at the workspaces[] array — that's the credential's effective allow-list.

Still stuck?

Open an issue at the package's npm page (link in the README) with:

• n8n version (docker exec <n8n> n8n -v)
• Node version
• Output of the Tellzm node's "Test Connection"
• Any relevant n8n container logs