• Resolved phlipineck

    (@phlipineck)


    “I’m using the HTTP Request step to call Firecrawl. The request succeeds, but I cannot retrieve the title from the response.

    I have tried:

    • {{ http_request1.data.metadata.title }}
    • {{ http_request1.data.title }}
    • {{ http_request1.metadata.title }}

    All fail with ‘required input title is missing’.

    The Log Message step shows {{ log1.logged }} as output, but I cannot see the actual response data.

    How do I correctly access the title from the Firecrawl response?

    Please provide the exact variable syntax for accessing nested JSON properties in HTTP responses.”

    The page I need help with: [log in to see the link]

Viewing 10 replies - 1 through 10 (of 10 total)
  • Thread Starter phlipineck

    (@phlipineck)

    The Firecrawl response structure (from their docs) is:

    {
    “success”: true,
    “data”: {
    “markdown”: “…”,
    “metadata”: {
    “title”: “…”
    }
    }
    }


    Thread Starter phlipineck

    (@phlipineck)

    Environment:

    • Meow Workflow: latest
    • WordPress: 7.1
    • PHP: 8.1
    Plugin Author Jordy Meow

    (@tigroumeow)

    Hi @phlipineck,

    You are very close. The HTTP Request step exposes three values: status (the HTTP code), body (the raw response text) and json (the body already parsed as JSON). Everything inside the response lives under that json value, so with the Firecrawl structure you posted the path is:

    {{ http_request1.json.data.metadata.title }}

    Your attempts were skipping the json part, which is why the value came back empty and the next step complained that its title was missing.

    Two small things that make this easier next time. After a test run, click the HTTP Request step: the panel on the right shows the full output of the last run (status, body and json), so you can see exactly what Firecrawl returned and work out the path from there. And the Log Message step only reports “logged” because the message itself goes to the log file (Settings β†’ For developers), not to the canvas, so the inspector is the quicker place to look at the response.

    Also double check the step ID shown in the inspector: it is http_request1 for the first HTTP step you add, but it becomes http_request2 if you added one before, or something else if you renamed it.

    Hope that gets your workflow running, and thanks for trying Meow Workflow this early! 😊

    Cheers,
    Jordy.

    Thread Starter phlipineck

    (@phlipineck)

    Hi Jordy,

    I’m still stuck on this. I have two problems:

    Problem 1: Webhook URL is blank

    • The workflow is published and set to “Active” (status bar says “Live up to date”)
    • But when I click on the Trigger step, the Webhook URL field is blank
    • There is no URL to copy

    Problem 2: Firecrawl returns “Invalid URL”

    • The Trigger step shows this sample payload:json{ “url”: “https://www.swotahtravel.com/golden-circle-trip” }
    • In the Firecrawl HTTP Request, the Body field is:json{ “url”: “{{ trigger.body.url }}”, “formats”: […] }
    • Firecrawl response shows:text”Invalid URL”
    • The output of the HTTP Request step shows null

    So the URL is in the sample payload, but it’s not being passed to Firecrawl correctly.

    Environment:

    • WordPress: 7.1
    • PHP: 8.1
    • Meow Workflow: latest

    Please help me fix this. I’ve spent many hours trying to get this to work.

    Thanks,
    Phil

    and….thank you for your time and for building this plugin!

    Plugin Author Jordy Meow

    (@tigroumeow)

    Hi Phil,

    Thank you for the details, and sorry for the hours lost. I reproduced both problems on a WordPress 7.1 site, and both are on my side, not yours.

    Problem 1, the blank Webhook URL. The URL is actually there. The WordPress admin styles give read-only fields a light grey background, and my field uses light text, so the URL ends up light on light and invisible. Version 0.1.3, released today, makes it visible again. Until you update, click inside the empty-looking field and press Ctrl+A then Ctrl+C (Cmd on Mac), or use the copy button on its right: the full URL is in the clipboard.

    Problem 2, the Firecrawl “Invalid URL”. The trigger hint in the plugin says the body is exposed as trigger.body, but the fields of your payload actually sit directly on trigger. So in the HTTP Request body, replace {{ trigger.body.url }} with:

    {{ trigger.url }}

    Firecrawl was receiving an empty url, hence the error and the null output. In 0.1.3 the hint is corrected, trigger.body works as well (so your current expression is fine too once updated), and the “required input is missing” error now tells you which reference came back empty. Please try the new 0.1.3 version.

    Once Firecrawl answers, the title is at {{ http_request1.json.data.metadata.title }} as mentioned before. If anything else gets in the way, tell me here and I will look at it right away.

    Thanks again for your patience, and for pushing Meow Workflow this far! 😊
    Jordy.

    Thread Starter phlipineck

    (@phlipineck)

    Hi Jordy,

    Just for your info the webhook url is still barely visible

    as you knwo I’m using Meow Workflow v0.1.3 on WordPress 7.1 with PHP 8.1.

    I have an HTTP Request step (step ID: gemini) that calls the Google Gemini API. The step’s output lists three variables: {{ gemini.status }}, {{ gemini.body }}, and {{ gemini.json }}.

    When I use {{ gemini.body }} in a later step, it correctly inserts the raw JSON response string.

    However, when I try to access the parsed JSON object using {{ gemini.json }} or a nested path like {{ gemini.json.candidates[0].content.parts[0].text }}, the output is empty.

    I also tried {{ gemini.candidates[0].content.parts[0].text }} (without .json), which was also empty.

    Could you please tell me the correct syntax to extract a nested value from the JSON response body? For example, how would I get the text field from:

    json{ “candidates”: [ { “content”: { “parts”: [ { “text”: “Rewritten content here…” } ] } } ] }

    Thank you!

    Phil

    Plugin Author Jordy Meow

    (@tigroumeow)

    Indexes are written with a dot, not brackets, so the path is:

    {{ gemini.json.candidates.0.content.parts.0.text }}

    That works in 0.1.3 right now. Brackets like candidates[0] were silently ignored, which is why everything came back empty. Version 0.1.4 (I have released it now) accepts the bracket form too, and after a test run the Insert data menu lists every nested field of the response with its value, so you can click the exact path instead of typing it. It also shows a step’s last output right at the top of the panel when you click it, so what an API returned is never hidden again.

    About {{ gemini.json }} on its own: in 0.1.3 a whole JSON object dropped into a text field comes out empty. 0.1.4 turns it into JSON text. Until then, {{ gemini.body }} gives you the same content as raw text.

    And thanks for the note on the webhook URL. If it is still faint after a hard refresh (Ctrl+Shift+R, or Cmd+Shift+R on Mac), the browser is holding the old script. In 0.1.4 the URL is shown as plain text in a light box that wraps on several lines, so nothing can hide it again.

    Thanks for testing all this so thoroughly, it is making the plugin better for everyone 😊

    Cheers,
    Jordy

    Thread Starter phlipineck

    (@phlipineck)

    Thanks for your help and patience Jordy – we got that working and thanks again for all your effort with this

    Phil

    Plugin Author Jordy Meow

    (@tigroumeow)

    You’re very welcome @phlipineck 😊 And feel free to leave a little review if you can, it really helps! No reviews means no downloads, and the more I get, the more motivated I am to keep working on it. Thank you!

    Thread Starter phlipineck

    (@phlipineck)

    Glowing feedback left!we are using in very simply to scrape a supplier webpage and use Gemini to rewrite for our own website – another project you might find interesting – we are having to use ‘Appearance’ and ‘Customize’ to write additional CSS from the Gutenberg block that Gemini delivers. I ‘block Conversion and Style’ app would be great.
    Thanks again going to look at you other apps to see what else we can use for our social, SEO and workflow

    Phil

Viewing 10 replies - 1 through 10 (of 10 total)

You must be logged in to reply to this topic.