Bubble
Bubble is a visual programming language and a PaaS developed by Bubble Group.
Verdict
Common use cases
- Look up customer accounts during support chats
- Update feature flags without opening the editor
- Pull user cohort data for product analysis
- Create test records for QA workflows
- Query app metrics for standup reports
Integration
- Vendor
- Bubble
- Category
- other
- Auth
- API_KEY
- Composio slug
bubble
Tools
Setup
Setup guide
- 11. In Switchy, open your workspace settings and navigate to the Integrations tab. 2. Click 'Add MCP Integration' and select Bubble from the list. 3. In a separate tab, log into your Bubble app editor, go to Settings > API, and enable the Data API if it's not already on. 4. Generate a new API key with read and write permissions for the data types you want Switchy to access. 5. Copy the API key and paste it into the Switchy connection form, then enter your Bubble app's subdomain (e.g. 'myapp' if your app is at myapp.bubbleapps.io). 6. Click 'Connect' and wait for Switchy to validate the credentials. 7. Open any Space, type '@Bubble list users' (replacing 'users' with one of your actual data type names), and confirm the MCP returns results.
What teammates see: by default, memories from Bubble are scoped to the Space (PROJECT visibility) - you can mark any memory PRIVATE or share it ORG-wide.
Works well with
Top models
Compatibility data appears once enough Spaces have used this MCP together with a given model.
How Switchy teams use it
Starter prompts
Fetch Recent Signups
@Bubble retrieve the 20 most recently created User records, sorted by creation date descending, and show me their email and signup source fields.Open in a Space →
Update Feature Access
@Bubble find the User record with email 'alex@example.com' and set the 'beta_access' field to true.Open in a Space →
Create Test Account
@Bubble create a new User record with email 'test+qa@example.com', name 'QA Tester', and role 'free'.Open in a Space →
Count Active Sessions
@Bubble count how many Session records have 'status' equal to 'active' and were created in the last 24 hours.Open in a Space →
Export Feedback Entries
@Bubble list all Feedback records submitted this week, including the user email, message text, and rating fields.Open in a Space →
Example outputs
Illustrative - representative of the model's voice and quality, not literal recordings.
@Bubble fetch the last 20 users who signed up this week and show their email domains
Retrieved 20 users created between Jan 13-19, 2025. Domain breakdown: 8 @gmail.com, 5 @company.com, 3 @outlook.com, 2 @protonmail.com, 2 other. Most recent signup was 14 minutes ago (user_1737294801). This example assumes a 'User' data type exists in your Bubble app with an email field and created_date.
Bubble's Data API requires you to specify the exact data type name from your app. Response structure mirrors your custom fields — if your User type has different field names, the output changes accordingly. Rate limits apply per your Bubble plan tier.
@Bubble create a new Project record with name 'Q1 Marketing Campaign' and status 'Planning'
Created Project record with unique_id proj_1737294923. Fields set: name = 'Q1 Marketing Campaign', status = 'Planning', created_date = 2025-01-19T15:42:03Z. This example assumes your Bubble app has a 'Project' data type with 'name' and 'status' fields defined. The record is now live in your database.
Write operations require API key permissions for the target data type. Field names must match your Bubble schema exactly — typos fail silently or return validation errors. Changes are immediate and irreversible through the API; consider workflow constraints in your Bubble app.
@Bubble pull all Orders with status 'pending' from the last 48 hours and calculate total revenue if they convert
Found 14 pending Orders created since Jan 17, 2025 15:42 UTC. Total potential revenue: $3,847.50 (assuming all convert at listed prices). Breakdown: 6 orders $100-$300 range, 5 orders $300-$500, 3 orders $500+. Highest single order: $890 (order_1737281004). This calculation uses the 'price' field from your Order data type.
The MCP retrieves raw data; the AI performs the revenue math. Accuracy depends on your Bubble schema having a numeric price field and consistent status values. If orders link to line-item tables, you'll need multiple queries or a Bubble API workflow to aggregate correctly.
Use-case deep-dives
When Bubble MCP makes sense for prototyping teams
A 3-person startup built their MVP in Bubble and now needs AI agents to read or write user records during customer onboarding flows. The Bubble MCP lets Claude query your app's database tables directly through API key auth, so you can automate tasks like flagging incomplete profiles or pulling usage stats into a daily Slack summary. This works best when your Bubble app is the single source of truth and you're comfortable exposing read/write access to an AI. If your data model changes weekly or you're syncing across multiple platforms, the MCP becomes a maintenance burden—you'll spend more time updating prompts than building features. For teams running stable Bubble apps with predictable schemas, the MCP turns your no-code backend into a first-class AI integration.
Pulling Bubble user context into support workflows
A 6-person SaaS team uses Bubble for their customer portal and wants support agents to see account history without switching tabs. The Bubble MCP connects Claude to your user database, so when a ticket comes in, the agent can ask 'show me this user's last 5 logins and subscription tier' and get an answer in seconds. This shortcut works when your support volume is under 50 tickets a day and your Bubble schema has clear field names—Claude can infer what to query. It breaks down if your data model uses cryptic internal IDs or if you need sub-second response times; the MCP adds a round-trip to Bubble's API, which can lag under load. For small teams doing tier-1 support on Bubble-backed products, the MCP cuts context-switching by half.
When Bubble MCP isn't the right call for analytics
A 4-person marketing team wants to segment Bubble app users by behavior and export lists for email campaigns. The Bubble MCP can technically query user tables and filter by signup date or feature usage, but it's not built for bulk exports or complex joins—you'll hit rate limits and slow responses if you're pulling thousands of records. Bubble's native CSV export or a direct Postgres connection (if you're on a paid plan) will be faster and more reliable. The MCP shines when you need one-off lookups or small batches (under 100 records), not when you're running weekly cohort analysis. If your segmentation logic changes every campaign, stick with Bubble's UI or a dedicated analytics tool.
Frequently asked
What does the Bubble MCP let me do in Switchy?
It connects your Bubble app's backend to Switchy so AI agents can read and write data in your no-code database. You can query records, update fields, trigger workflows, and pull app metadata without switching contexts. Think of it as giving your team's AI direct access to your Bubble data tables and API workflows.
Do I need admin access to connect Bubble?
Yes. You'll generate an API key from your Bubble app's Settings > API tab, which requires admin or developer permissions. The key grants full read-write access to your app's data and workflows, so only connect accounts you trust. Standard Bubble collaborators can't create API keys.
Can the MCP trigger Bubble workflows or just read data?
It can do both. If your Bubble app exposes backend workflows via the API, the MCP can trigger them with parameters. It can also read from your database tables and modify records if your API settings allow it. Check your Bubble app's API settings to confirm which workflows and data types are exposed.
Why use this instead of Bubble's built-in API?
Bubble's API requires you to write and test HTTP requests manually. The MCP wraps that complexity so your AI agents can interact with Bubble data in natural language. You skip the boilerplate of constructing JSON payloads and parsing responses — the agent just asks for what it needs.
Who on my team should connect the Bubble MCP?
Whoever owns your Bubble app's backend and understands which data types and workflows are safe to expose. This person should audit the API key's scope before connecting it to Switchy. Once connected, any Switchy workspace member can use the MCP in their prompts.