Troubleshooting

Almost every issue has a clear error message in Settings → ReplyMind → Logs. Open that tab first — find the most recent failed entry and read the message in the AI Reply column. The text below covers the most common ones.

First step: check the Logs tab

Almost every issue has a clear error message in Settings → ReplyMind → Logs. Open that tab first — find the most recent failed entry and read the message in the AI Reply column. The text below covers the most common ones.

Errors in the AI Reply column

"API key is not configured"

Cause: No API key saved.

Fix: Go to Settings → ReplyMind, paste your API key, click Save Changes.

"Missing API key."

Cause: Same as above — the API key was deleted, or saving failed.

Fix: Re-enter the key. If saving doesn't persist, check that you're an Administrator (capability manage_options is required to save settings).

"AI service returned HTTP 401"

Cause: The API key is rejected — either invalid, revoked, or the wrong provider for the key.

Fix:

  1. Confirm the key is correct on the provider's dashboard (OpenAI / Anthropic).
  2. Confirm the Provider dropdown matches the key (sk- keys → OpenAI; sk-ant- keys → Anthropic Claude).
  3. If you suspect the key was leaked, regenerate it on the provider's dashboard, then click Remove key in the plugin and paste the new one.

"AI service returned HTTP 429"

Cause: Rate-limited by your AI provider — usually too many requests in a short window.

Fix:

  • Wait 1–2 minutes and click Generate AI Reply to retry.
  • If it happens often, your provider's rate-limit tier is too low — upgrade your plan with the provider, or use a cheaper model that allows higher RPM.

"AI service returned HTTP 400"

Cause: The request was malformed — usually because a model name was changed or the prompt is too long.

Fix:

  1. Re-select a Model from the dropdown and save.
  2. If your custom System Prompt is very long (multiple paragraphs), shorten it.

"AI service returned HTTP 500" or "HTTP 503"

Cause: The AI provider is having an outage.

Fix: Wait 5–10 minutes and retry. Check the provider's status page (OpenAI Status / Anthropic Status).

"Daily AI reply limit reached"

Cause: The plugin's daily cap (default 100) has been hit.

Fix: Either wait until midnight (server timezone), or raise the cap:

add_filter( 'replymind_daily_limit', fn() => 500 );
// Set to 0 to disable the cap entirely

Drop into your theme's functions.php or a custom plugin.

"Unable to parse AI response."

Cause: The AI returned something the plugin couldn't read — usually empty content, or a content-policy block from the provider.

Fix: Click Generate AI Reply to retry. If it keeps failing on the same comment, the comment text might be triggering a content filter at the provider — exclude it via the replymind_send_comment_data filter or skip it manually.

"AI service returned an empty response."

Cause: The provider returned an empty completion.

Fix: Retry. If persistent, switch to a different model.

Drafts never appear

If new comments come in but nothing happens, the WP-Cron isn't firing.

Why

The plugin uses wp_schedule_single_event to defer the AI call so the visitor's submission isn't blocked. WP-Cron normally runs on page visits — on a low-traffic site, hours can go by without it triggering.

Fix

Set up a real cron job to hit wp-cron.php every 5 minutes. Most hosts have a one-click toggle:

  • cPanel / Plesk — look for "Cron Jobs" and add: */5 * * * * curl -s https://yoursite.com/wp-cron.php?doing_wp_cron > /dev/null
  • Cloudways / Kinsta / WP Engine — these typically have a built-in cron control in their dashboard.

Optionally, disable WP-Cron's page-visit fallback in wp-config.php:

define( 'DISABLE_WP_CRON', true );

Replies post immediately even though I'm in Suggestion mode

Cause: The plugin is configured for auto mode, not suggestion.

Fix: Settings → ReplyMind → Reply Mode dropdown → set to Suggestion Mode → Save Changes.

Verify the change took effect by submitting a test comment.

"Provider/model mismatch" warning

Message: "Provider is set to OpenAI but a Claude model is selected" (or vice versa).

Cause: You changed the provider but the model dropdown wasn't updated.

Fix: Settings → ReplyMind → re-select a Model that matches your Provider → Save.

The Generate AI Reply link doesn't appear in Comments

Cause: One of:

  • The comment is in spam / trash / pending — the plugin only acts on approved comments.
  • An AI reply already exists for that comment (_replymind_reply_generated is set).
  • Your role doesn't have the edit_comment capability (typically Editor or Administrator).

Fix: Approve the comment, or check your user role.

Activation redirect loops

If activating the plugin keeps re-triggering the redirect to Settings:

Fix: Manually delete the option:

delete_option( 'replymind_do_activation_redirect' );

Run via WP-CLI:

wp option delete replymind_do_activation_redirect

Plugin "missing" headers in Plugin Check

If Plugin Check reports the plugin's Description / Version / License as missing, you're scanning the wrong plugin or the wrong file. The headers in replymind-ai-comment-responder.php are valid. Re-select the correct slug in Plugin Check's dropdown and re-run.

Still stuck?

When asking for help, paste the most recent error from Settings → ReplyMind → Logs along with your WordPress version, PHP version, and the AI provider you're using.