# 🔧 \[SOLVED\] How I received, modified, and resent a large JSON in Bubble using Toolbox & Backend Workflows **URL:** \ **Category:** Content / Articles\ **Created:** [May 19, 2025, 10:32am UTC](https://forum.bubble.io/t/solved-how-i-received-modified-and-resent-a-large-json-in-bubble-using-toolbox-backend-workflows/370077 "2025-05-19T10:32:41Z")\ **Posts on this page:** 3\ **Page:** 1
**Author:** ![votterskog](https://sea1.discourse-cdn.com/bubble/user_avatar/forum.bubble.io/votterskog/32/243952_2.png) [@votterskog](https://forum.bubble.io/u/votterskog)\ **Post date:** [May 19, 2025, 10:32am UTC](https://forum.bubble.io/t/solved-how-i-received-modified-and-resent-a-large-json-in-bubble-using-toolbox-backend-workflows/370077/1 "2025-05-19T10:32:41Z")
Hey Bubble community! I wanted to share a solution I finally cracked after 30+ hours of trial and error — in case it helps someone else out there struggling with JSON modification and external API workflows. 🧩 My challenge I needed to: 1. **Receive structured JSON content** from an external API (in my case, Eventbrite), 2. **Modify** it — specifically update an image inside a nested object and bump the `page_version_number`, and 3. **Send** the modified version back to Eventbrite. ❌ What didn’t work I initially set the API Connector to expect **JSON** as the response type. The problem? Bubble wouldn’t let me easily work with or re-use the entire JSON structure — especially when dealing with complex nested arrays and objects. Trying to pass this into toolbox (server script) didn’t work. ✅ The fix: Treat the response as **text** and process it manually Here’s the key insight: - **Set the response type to “text”** in the API Connector. ![CleanShot 2025-05-19 at 11.31.27@2x](https://us1.discourse-cdn.com/bubble/original/3X/4/e/4e60bf874db7c15da82d8f4020c2fda97a559f5a.png) - Save the response as a plain text field in your database or pass it straight into server script. - Use **Toolbox’s “Run JavaScript”** in a backend workflow to parse, modify, and rebuild the JSON. - Send the final result wherever you need — re-encoded as JSON. 🔍 The JavaScript I used in Toolbox (server-side action) ```auto async function updateJson() { try { let json = JSON.parse(properties.thing1); let new_image_url = properties.thing2; let new_image_id = properties.thing3; // ✅ Update herocarousel image if (json.widgets && json.widgets[0]?.data?.slides[0]?.image) { json.widgets[0].data.slides[0].image.url = new_image_url; json.widgets[0].data.slides[0].image.image_id = new_image_id; } // ✅ Clean function function clean(obj) { if (Array.isArray(obj)) return obj.map(clean); if (obj && typeof obj === 'object') { const cleaned = {}; for (const key in obj) { if ( key === "resource_uris" || key === "resource_url" || key === "semantic_purpose" || key === "timed_display" || key === "display_restrictions" || (key === "id" && obj[key] === "") ) continue; cleaned[key] = clean(obj[key]); } return cleaned; } return obj; } // ✅ Keep only these widget types, and remove duplicate 'faq' if both exist const allowedWidgets = []; const widgetTypes = new Set(); for (const w of json.widgets || []) { if (["herocarousel", "faqs"].includes(w.type) && !widgetTypes.has(w.type)) { allowedWidgets.push(w); widgetTypes.add(w.type); } } const cleanedJson = { access_type: json.access_type, purpose: json.purpose, modules: clean(json.modules), widgets: clean(allowedWidgets) }; return JSON.stringify(cleanedJson); } catch (error) { return JSON.stringify({ error: error.message }); } } updateJson(); ``` ![CleanShot 2025-05-20 at 04.39.13@2x](https://us1.discourse-cdn.com/bubble/original/3X/4/c/4cfc3ae5b695081fdf8c843fb27b9e36a5564cb2.png) Why this works better - By treating the original API response as **raw text** , you sidestep Bubble’s sometimes-too-strict JSON parsing. - Toolbox gives you full JS power — no more nested “:extract with Regex” madness. - Your backend stays clean, scalable, and debuggable. ![CleanShot 2025-05-20 at 04.35.37@2x](https://us1.discourse-cdn.com/bubble/original/3X/e/6/e66778c2829413784872dd27811c19cebc25b511.png) I had to do a normal GET request to also be able to bump the version number. Other than that I posted the JSON body as a raw body back to eventbrite. All works fine. **Bubble’s API Connector must receive structured content as text** , not JSON, if you want to parse and manipulate it. At least from the Eventbrite API as far as I can work it. I tried sending it off to AWS Lambda, Make, and even Zerowork — nothing worked. It wasn’t until I discovered that I had to receive the JSON as raw **text** in the API Connector that things finally clicked. Thanks for reading, I hope it can help you with something. ---
**Author:** ![randomanon](https://avatars.discourse-cdn.com/v4/letter/r/8491ac/32.png) [@randomanon](https://forum.bubble.io/u/randomanon)\ **Post date:** [May 20, 2025, 2:32am UTC](https://forum.bubble.io/t/solved-how-i-received-modified-and-resent-a-large-json-in-bubble-using-toolbox-backend-workflows/370077/2 "2025-05-20T02:32:28Z")
Let’s say you have a triple nested JSON: - Countries – States — Cities Do you create a dummy JSON response first then receive it as that data type? How would you actually display that in your elements? ---
**Author:** ![votterskog](https://sea1.discourse-cdn.com/bubble/user_avatar/forum.bubble.io/votterskog/32/243952_2.png) [@votterskog](https://forum.bubble.io/u/votterskog)\ **Post date:** [May 20, 2025, 3:25am UTC](https://forum.bubble.io/t/solved-how-i-received-modified-and-resent-a-large-json-in-bubble-using-toolbox-backend-workflows/370077/3 "2025-05-20T03:25:40Z")
If I understand you correctly. - In your API connector use as data. - Do a get request. - On the element I think you’d choose “get data from an external API” response^country^state’s^city