Skip to content

troubleshooting.md

Latest commit

 

History

History
76 lines (58 loc) · 2.87 KB

troubleshooting.md

File metadata and controls

76 lines (58 loc) · 2.87 KB
Raw

Troubleshooting

Connector / setup

The connector doesn't appear after I add it. Refresh Claude (reload the tab or restart the desktop app). On Team/Enterprise, the connector must be added by an Owner in Organization settings → Connectors before members can see it under Customize → Connectors.

The browser auth window opens, then closes, and nothing is connected. Pop‑up blockers and third‑party‑cookie blockers are the usual cause. Allow pop‑ups for claude.ai and apify.com, then click Connect again. If you are signed in to more than one Apify account, sign out of the others first.

"No tools available" even though sign‑in succeeded. Your ?tools= value may be empty or malformed. Edit the connector and set it to exactly:

https://mcp.apify.com/?tools=actors,docs

Reconnect if prompted.

I'm on the Claude Free plan and can't add it. Free plans allow one custom connector. Remove an existing one, or upgrade.

I want my own API token instead of OAuth. Only some Claude Desktop configurations accept a JSON block with an Authorization: Bearer header — see connect-claude-mcp.md. On claude.ai, use OAuth.

Runs

Claude says it can't find the Actor. Use the full name in your request: "the Orgupdate Reed jobs scraper" or the slug orgupdate/reed-jobs-scraper. Check the slug against actors.md. If you pinned Actors in the URL, make sure the slug is spelled correctly there.

The run returns no results.

  • Confirm includeKeyword is not empty.
  • Use a job‑title style query, not a long descriptive phrase.
  • Check the spelling of locationName and countryName.
  • Remove jobType / datePosted filters and try again.
  • Increase pagesToFetch. You are not charged the per‑result fee for an empty run.

Fewer results than I expected. Some queries genuinely have thin coverage on a given board. Broaden the keyword ("engineer" instead of "senior staff backend engineer"), widen datePosted, raise pagesToFetch, or run a second source.

"Budget exceeded" / the run stops early. Raise the run budget limit in Apify (Console → the run → Budget), or lower pagesToFetch. Pay‑per‑result runs need enough budget for at least one page.

Duplicate jobs across runs or sources. The same posting often appears on several boards. Ask Claude to "remove duplicates by URL, then by company + title". For scheduled scraping set datePosted to today or 3days so each run only returns new postings.

Some listings are missing salary / job type / exact date. Normal — not every employer publishes those fields. Tell Claude how to handle blanks in your output.

Still stuck?

  • Check the Actor's own page on the Apify Store for status and notes.
  • Open an issue on this repository with the request you sent Claude and what came back.