developer-toolsoauth2

DevCycle

DevCycle is a feature management platform for creating, targeting, releasing, and monitoring feature flags across applications.

Verdict

DevCycle MCP lets your team manage feature flags, variables, and targeting rules without leaving the AI workspace. @mention it to create typed variables, inspect feature configurations across environments, list audiences, or audit which flags are live. Engineers shipping incremental rollouts get instant visibility into what's enabled where; product managers can verify targeting logic before a launch. The OAuth flow requires project-level write access, so you'll grant broader permissions than a read-only audit would need.

Common use cases

  • Audit which feature flags are live in production
  • Create typed variables for new experiments
  • Compare targeting rules across staging and prod
  • List audiences used in active campaigns
  • Delete deprecated flags after cleanup

Integration

Vendor
DevCycle
Category
developer-tools
Auth
OAUTH2
Tools
9
Composio slug
dev_cycle

Tools

  • Create Variable

    Create a typed variable in a DevCycle project, optionally associating it with a feature.

  • Delete Variable
    destructive

    Permanently delete exactly one DevCycle variable, selected by key or stable ID, from the specified project. This operation cannot be undone.

  • List Audiences

    List and search project audiences, optionally including whether each audience is used.

  • List Environments

    List environments in a DevCycle project without exposing SDK key values.

  • List Feature Configurations

    Inspect a feature's targeting configurations across all environments or one selected environment.

  • List Features

    List and filter current Feature v2 resources in a DevCycle project.

  • List Projects

    List DevCycle projects visible to the connected Management service account, with filtering and controlled pagination.

  • List Variables

    List and filter variables in a DevCycle project, including variables associated with a feature.

  • Update Variable

    Update one or more editable properties of an existing DevCycle variable.

Setup

Setup guide

  1. 11. Inside any Switchy Project, type '@' and select 'Connect new integration' from the menu. 2. Choose DevCycle from the developer tools category. 3. Click 'Connect with DevCycle' to start the OAuth flow in a new browser tab. 4. Log in to your DevCycle account and select the organization you want to connect. 5. Grant Switchy permission to read projects, environments, features, variables, and audiences, plus write access to create or delete variables. 6. After authorizing, the tab closes and you return to Switchy with DevCycle connected. 7. Test the connection by typing '@DevCycle list projects' in the chat — you should see your DevCycle projects listed in the response. 8. To invoke a tool in a Project, @mention DevCycle followed by your request, like '@DevCycle show me feature configurations for checkout-v2 in production'.

What teammates see: by default, memories from DevCycle are scoped to the Project (PROJECT visibility) - you can mark any memory PRIVATE or share it ORG-wide.

Works well with

Top models

Compatibility data appears once enough Projects have used this MCP together with a given model.

How Switchy teams use it

Not enough Projects yet to publish anonymised usage stats (we require ≥ 50 Projects per week).

Starter prompts

Audit Production Flags

@DevCycle list all features in the production environment and show me which ones are currently enabled

Create Experiment Variable

@DevCycle create a boolean variable called 'new-checkout-flow' in the main project and associate it with the checkout-redesign feature

Compare Environment Configs

@DevCycle show me the feature configurations for 'premium-upsell' in both staging and production environments

List Active Audiences

@DevCycle list all audiences in the project and tell me which ones are currently being used

Clean Up Old Variables

@DevCycle delete the variable 'legacy-payment-gateway' from the main project

Example outputs

Illustrative - representative of the model's voice and quality, not literal recordings.

Prompt

@DevCycle list all features in the mobile-app project and show which environments they're enabled in

Output

Found 7 features in mobile-app. Three are live in production: 'new-checkout-flow' (enabled for 100% of users), 'dark-mode' (enabled for beta audience), and 'push-notifications' (enabled for users in US region). Two features are staging-only: 'experimental-search' and 'redesigned-profile'. Two features are disabled across all environments: 'legacy-payment-ui' and 'social-sharing'. Each feature includes its current targeting rules and rollout percentages per environment.

Notes

This read operation requires OAuth scope to view project features. The response reflects the state at query time—feature flags may change between queries if your team is actively deploying. Useful for auditing what's live before a release or debugging unexpected behavior in specific environments.

Prompt

@DevCycle create a boolean variable called 'enable_ai_suggestions' in the web-platform project and associate it with the smart-editor feature

Output

Created variable 'enable_ai_suggestions' (type: Boolean, default: false) in web-platform project. Variable is now associated with feature 'smart-editor' and available for targeting across all environments. You can configure per-environment values and targeting rules in the DevCycle dashboard or by updating feature configurations through this MCP. Variable key is stable and can be referenced in your application code immediately.

Notes

This write operation permanently adds a variable to your project. Requires OAuth scope for variable creation. The variable exists but won't affect application behavior until you configure targeting rules in at least one environment. Consider setting a safe default value (typically false for new features) to avoid unintended rollouts.

Prompt

@DevCycle check the targeting configuration for the 'premium-analytics' feature across all environments and summarize which user segments will see it

Output

The 'premium-analytics' feature uses different targeting strategies per environment. In development: enabled for all internal team members (audience: 'devcycle-internal'). In staging: enabled for 25% of users with email domain '@company.com' plus the 'beta-testers' audience (estimated 340 users). In production: enabled only for users with account_type='enterprise' AND region in ['US', 'CA', 'UK'] (estimated 2,100 users, 8% of total traffic). No targeting configured in the 'preview' environment—feature is fully disabled there.

Notes

This synthesis combines feature configuration data with AI reasoning to translate technical targeting rules into plain-language summaries. Useful for release planning or explaining rollout strategy to non-technical stakeholders. Audience size estimates depend on DevCycle's analytics data availability and may lag real-time user counts by several minutes.

Use-case deep-dives

Feature rollout coordination across environments

When DevCycle wins for multi-environment feature releases

A 6-person engineering team shipping a new checkout flow needs to coordinate feature flags across staging, canary, and production. DevCycle's List Feature Configurations and List Environments tools let the team audit targeting rules in one pass during their release standup. The OAuth2 flow means each engineer authenticates once, then the shared workspace shows real-time flag state without passing SDK keys around Slack. This works cleanly when you have 3-8 environments and ship weekly or faster. If you only toggle flags once a quarter or run a single production environment, the MCP overhead outweighs opening the DevCycle dashboard. The buying call: teams running continuous deployment with environment parity use this to catch misconfigured rollouts before they hit users.

Customer support flag lookup

Using DevCycle MCP for live customer troubleshooting

A 3-person support team fields tickets about inconsistent UI behavior tied to feature flags. The List Variables and List Audiences tools let them search which flags apply to a specific customer segment without asking engineering to check the dashboard. The MCP surfaces variable types and audience membership in the shared Switchy workspace, so the support rep can confirm "yes, enterprise users see the new nav" in under 30 seconds. This scenario breaks down if your flag taxonomy exceeds 100 variables or you need to modify targeting on the fly—DevCycle MCP is read-heavy, and the Create/Delete Variable tools require project-write permissions most support teams shouldn't have. The threshold: if you troubleshoot flags more than twice a week, this saves 15 minutes per incident.

Audit trail for compliance reporting

When DevCycle MCP handles feature flag compliance checks

A fintech startup's 4-person ops team needs quarterly proof that payment-related features were gated correctly across regions. The List Features and List Feature Configurations tools pull targeting history into the Switchy workspace, where the team exports a snapshot showing which audiences saw which variables in each environment. The OAuth2 auth means the audit log ties back to named users, not shared API keys. This works when you have 10-40 features under regulatory scope and audit once a quarter or month. If you need real-time compliance monitoring or your feature count exceeds 200, you're better off piping DevCycle webhooks to a dedicated audit database. The call: teams with light compliance needs use this to avoid building custom reporting infrastructure.

Frequently asked

What does the DevCycle MCP let me do in Switchy?

The DevCycle MCP connects your feature flag infrastructure to Switchy's AI workspace. You can list projects, environments, and features, inspect targeting configurations, and create or delete variables without leaving your chat. It's useful when you need to audit flag state, coordinate releases across teams, or quickly spin up new variables during planning sessions.

What OAuth permissions does DevCycle ask for when I connect it?

DevCycle uses OAuth2 to authenticate your Management API account. The specific scopes depend on your DevCycle plan and role, but you'll need read access to list projects and features, plus write access if you want to create or delete variables. Connect with an account that has project-level permissions for the environments your team works in.

Can the MCP change feature flag targeting rules or toggle flags on and off?

No. The MCP lets you inspect targeting configurations and list which features exist, but it doesn't modify targeting rules or flip flags between on and off states. For runtime control, use DevCycle's dashboard or SDK directly. The MCP is read-heavy by design to prevent accidental production changes during casual AI conversations.

Why use this instead of just logging into DevCycle's web console?

The MCP shines when you're already working in Switchy and need to cross-reference flag state with code, tickets, or deployment logs. Instead of context-switching to the DevCycle UI, you ask the AI to list features or check variable associations inline. It's faster for audits and planning, not a replacement for the full dashboard.

Who on the team should connect the DevCycle integration?

Connect it with a service account or team lead who has read access to all projects your team ships. If you need to create or delete variables from Switchy, that account must have write permissions too. Each connected account counts as one seat in your Switchy workspace, so share the connection rather than having everyone authenticate separately.

More developer tools MCPs

See all developer tools MCPs →

Data last verified 61 hours ago.Sources aggregated hourly to weekly. See docs/architecture/model-directory.md.