OpsLevel MCP Server – Developer Portal Context
The official OpsLevel MCP Server gives AI tools read-only access to an OpsLevel account through the Model Context Protocol. It is useful when an assistant needs service ownership, component, documentation, repository, dependency, system, team, user, infrastructure, check, campaign, or action context from an internal developer portal.
Overview
OpsLevel MCP Server is the official Model Context Protocol server from OpsLevel for connecting AI assistants to an OpsLevel account. OpsLevel is an internal developer portal and service catalog, so the MCP server is designed to give an AI agent structured, account-specific context about software ownership, components, documentation, standards, dependencies, infrastructure, systems, and teams.
The server is most useful when a developer or platform engineer wants an AI assistant to answer questions from the live OpsLevel catalog instead of relying on static model knowledge. For example, an agent can help identify who owns a service, find related components, inspect dependencies, review service documentation, or understand catalog metadata while planning engineering work.
What the MCP server enables
The official repository describes the server as read-only. It can read OpsLevel resources such as Actions, Campaigns, Checks, Components, component dependencies and dependents, API and Tech Docs, Domains, Filters, Infrastructure, Repositories, Systems, Teams, and Users. That makes it well suited for context-gathering and discovery workflows, not for making changes to the OpsLevel account.
Because the server is read-only, agents should use it to retrieve and summarize data rather than to create, update, or delete records. Access is still governed by the OpsLevel API token used to run the server, so teams should provision the token with appropriate permissions and rotate it according to normal secret-management practices.
When to use it
Use OpsLevel MCP Server when an AI assistant needs engineering-system context from the service catalog. Practical examples include asking which team owns a component, finding docs attached to a service, mapping dependencies before a change, identifying standards or checks that apply to a component, or discovering the repositories and systems related to a service.
Connection and authentication
OpsLevel documents a local stdio MCP setup. The server can be installed with Homebrew, run as a downloaded binary from GitHub releases, or run through Docker using the published public.ecr.aws/opslevel/mcp:latest image. The documented client examples use the opslevel-mcp command and authenticate with the OPSLEVEL_API_TOKEN environment variable.
Key considerations
The current server is intended for read-only account access. Tools require a valid OpsLevel API token; recent release notes indicate the server may start without a token for inspection and testing, but tools will not work until a valid token is provided. For production use, keep the token out of checked-in configuration files and use your MCP client's secret-input or environment-variable mechanism.
Supported Transports
stdio
Command: opslevel-mcp
Frequently Asked Questions
- When should an AI agent use the OpsLevel MCP Server?
- Use it when the agent needs live OpsLevel developer-portal context, such as service ownership, component metadata, dependencies, documentation, repositories, systems, teams, users, checks, campaigns, or infrastructure records.
- What does the OpsLevel MCP Server add to an AI agent's capabilities?
- It lets the agent query the organization's OpsLevel account through MCP so answers can be grounded in catalog data instead of only general model knowledge or manually pasted context.
- What can the AI agent access or manage through this server?
- The official server currently provides read-only access to OpsLevel resources including Actions, Campaigns, Checks, Components, dependencies, dependents, Documentation, Domains, Filters, Infrastructure, Repositories, Systems, Teams, and Users. It is not documented as a write or administration server.
- How is authentication configured for this MCP server?
- The documented setup uses an OpsLevel API token supplied through the OPSLEVEL_API_TOKEN environment variable. Store the token in the MCP client's secret mechanism or environment, do not hard-code it in shared configuration, and scope or rotate it according to your organization's practices.
- Which transport should be used for the OpsLevel MCP Server?
- Use the local stdio transport documented by OpsLevel client examples. The server can be launched as the opslevel-mcp binary, or via Docker with the OpsLevel image if the local binary is not installed.