AppDrag
AppDrag is a cloud-based platform for building websites, APIs, and databases with drag-and-drop tools, code editing, and integrated hosting to accelerate development workflows and iteration
Verdict
Common use cases
- Query production database row counts mid-standup
- Check serverless function logs after deploy
- Pull user signup trends for weekly review
- Verify API endpoint status during incident
- List recent schema changes before sprint planning
Integration
- Vendor
- AppDrag
- Category
- developer-tools
- Auth
- API_KEY
- Composio slug
appdrag
Tools
Setup
Setup guide
- 11. Open your Switchy workspace and navigate to Settings > Integrations > Developer Tools. 2. Click 'Connect AppDrag' and select API Key as the authentication method. 3. In a separate tab, log into your AppDrag dashboard, go to Account Settings > API Keys, and generate a new key with read access to the databases and functions you want Switchy to query. 4. Copy the key and paste it into the Switchy connection dialog, then click Authorize. 5. Switchy will verify the key and list the AppDrag projects now accessible. 6. To test, open any Space and type '@AppDrag list my databases' — you should see a response with your project's database names. 7. If the connection fails, double-check the key hasn't expired and that your AppDrag account has active projects.
What teammates see: by default, memories from AppDrag 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
Database Schema Overview
@AppDrag show me the schema for the users table in my production database, including column types and any indexes.Open in a Space →
Recent Function Executions
@AppDrag list the last 10 executions of my 'sendWelcomeEmail' function with timestamps and any error messages.Open in a Space →
User Growth This Month
@AppDrag how many new users signed up in my app this month, grouped by week?Open in a Space →
API Endpoint Health Check
@AppDrag check the status of my /api/orders endpoint and show me the last response time and status code.Open in a Space →
Deployment Log Summary
@AppDrag summarize the last 5 deployments to production, including commit hashes and timestamps.Open in a Space →
Example outputs
Illustrative - representative of the model's voice and quality, not literal recordings.
@AppDrag list all databases in my project and show their current storage usage
Based on your AppDrag project configuration, you have three databases: 'prod_users' (PostgreSQL, 2.3 GB used), 'staging_analytics' (MySQL, 847 MB used), and 'dev_sandbox' (PostgreSQL, 124 MB used). The prod_users database is approaching 80% of its allocated tier limit. All databases are currently online with no reported connection issues.
This example demonstrates read-only project inspection. Requires an API key with database:read scope. The MCP surfaces infrastructure state that's otherwise buried in the AppDrag dashboard, useful for capacity planning during standups.
@AppDrag deploy the staging branch to my preview environment and confirm when live
Deployment initiated for branch 'staging' to environment 'preview-env-01'. Build completed in 43 seconds with no errors. The preview environment is now live at preview-env-01.appdrag.site. Previous deployment (commit a3f2c91) has been archived. Note that environment variables were inherited from the last successful deploy.
This is a write action that triggers a real deployment. Requires deploy:write scope. The MCP does not roll back automatically on failure — you must monitor build logs separately. Useful for shipping preview links to stakeholders without leaving the conversation.
@AppDrag review the last 5 API error logs from production and suggest what might be causing the 502s
The last five production errors show four 502 Bad Gateway responses and one timeout. All occurred between 14:32 and 14:38 UTC, targeting the /api/v2/orders endpoint. The errors coincide with a spike in database connection pool exhaustion (max connections reached). This suggests either a sudden traffic surge or a slow query blocking the pool. Check for unindexed queries on the orders table or consider scaling your database tier if traffic is sustained.
This example pairs AppDrag's log retrieval with the AI's diagnostic reasoning. Requires logs:read scope. The AI's suggestions are based on patterns in the returned data, not live profiling — always verify with your own monitoring tools before making infrastructure changes.
Use-case deep-dives
When AppDrag fits rapid front-end iteration cycles
A 3-person design studio ships client mockups as live prototypes instead of static PDFs. AppDrag's API-key auth means the Switchy workspace can trigger deployments without OAuth ceremony, and the team avoids context-switching to a separate CI dashboard. This works when the prototype is front-end only—HTML, CSS, light JavaScript—and the client needs a URL within hours, not days. The trade-off: if your prototype needs backend logic beyond basic forms, AppDrag's tooling thins out fast and you'll hit friction around database schemas or serverless functions. Best fit is studios doing 5-10 client prototypes a month where speed trumps architectural depth. If that's your cadence, the MCP saves the 20-minute deploy dance every handoff.
AppDrag MCP for non-technical campaign launches
A 6-person growth team runs A/B tests on landing pages for SaaS trials. The marketer writes copy in Notion, the designer tweaks a Figma comp, and the Switchy agent pushes the new page live via AppDrag's API without waiting on engineering. API-key auth means the marketing lead controls access, not IT. This scenario wins when pages are standalone—no shared navigation, no CMS dependencies—and the team ships 2-4 variants a week. The boundary: if your landing pages pull dynamic pricing from Stripe or user testimonials from a database, AppDrag's static-first model becomes a bottleneck and you'll need a proper headless CMS. For pure marketing velocity on isolated pages, the MCP cuts deploy time from 2 hours to 10 minutes.
When AppDrag handles low-stakes internal dashboards
A 12-person operations team builds a shift-scheduling dashboard that reads from Google Sheets and displays who's on call. The dashboard is a single HTML file with a JavaScript fetch call—no build step, no framework. AppDrag's MCP lets the ops lead update the page from Switchy when the on-call rotation changes, and API-key auth keeps it inside the ops Slack channel without involving security review. This works when the tool is read-only or writes back to a third-party API, and uptime expectations are "business hours, best effort." If the dashboard needs to handle PII, require SSO, or guarantee 99.9% uptime, AppDrag's infrastructure isn't hardened for that and you'll need a real app platform. For throwaway internal tools updated monthly, the MCP is the right weight class.
Frequently asked
What does the AppDrag MCP do in Switchy?
The AppDrag MCP connects your AppDrag projects to Switchy's AI workspace, letting your team query project data, manage cloud functions, and interact with your AppDrag databases through natural language. Since AppDrag handles both backend and frontend infrastructure, the MCP gives your AI agents direct access to deployment logs, API endpoints, and database schemas without switching contexts.
Do I need an AppDrag admin account to set up the API key?
Yes. You'll need owner or admin permissions in your AppDrag account to generate an API key with sufficient scope. Standard collaborator accounts typically can't create keys that access project settings or database credentials. If you're on a team plan, ask whoever controls billing to generate the key and share it through Switchy's secure credential store.
Can the MCP deploy code changes or only read project data?
That depends on the API key permissions you configure in AppDrag. Most teams start with read-only access for safety — letting AI agents fetch logs, query databases, and inspect cloud functions without risk. If you explicitly grant write permissions when generating the key, agents could trigger deployments or modify database records. Start narrow and expand scope as you trust the workflow.
Why use this instead of just logging into AppDrag directly?
The MCP lets your entire team ask questions about AppDrag projects in Switchy's shared chat without each person needing AppDrag credentials or knowing where to find specific logs. An agent can pull error traces from three different cloud functions, correlate them with database query times, and summarize the issue in one response — work that normally requires clicking through multiple AppDrag dashboards.
Who on the team should connect the AppDrag integration?
Your DevOps lead or whoever manages your AppDrag account should connect it. They understand which projects contain sensitive data and can scope the API key appropriately. Once connected in Switchy, any team member with workspace access can query AppDrag through AI agents, but the original connector controls whether the key allows read-only or write operations.