currents-dev/currents-mcp

Overview

Currents MCP Server provides a bridge between AI assistants and the Currents test‑automation platform, enabling agents to ingest real‑time CI data and historical performance metrics. By exposing Currents’ rich API through the Model Context Protocol, developers can ask AI models to analyze failing tests, suggest optimizations, or generate remediation steps without leaving their IDE. The server translates MCP tool calls into authenticated Currents requests, returning structured JSON that AI assistants can readily consume.

The core value lies in turning opaque test logs into actionable insights. When a CI pipeline reports a flake or regression, an AI agent can invoke to retrieve the exact stack trace and environment snapshot, then call to inspect execution traces of the failing spec. The agent can further compare current run performance against historical data via and , automatically flagging performance regressions or flaky behavior. This tight integration removes the need for developers to manually sift through dashboards, streamlining debugging and reducing time‑to‑fix.

Key capabilities include:

• Project discovery () to enumerate all test projects in an organization.
• Run enumeration () for quick access to recent executions.
• Detailed run data () that provides metadata such as commit hash, branch, and runtime environment.
• Spec‑level diagnostics () exposing step‑by‑step execution data and failures.
• Performance analytics (, ) to surface slow or flaky tests over time.
• Signature retrieval () and result extraction () for pinpointing specific test artifacts.

Real‑world use cases span continuous integration monitoring, automated incident response, and proactive performance tuning. For example, a CI system can trigger an AI workflow whenever a test fails; the assistant queries Currents for context, proposes code changes or configuration tweaks, and even drafts a pull request. In larger teams, the server enables shared knowledge bases where AI can surface recurring patterns across multiple projects, fostering collective debugging practices.

Integration is straightforward for any MCP‑compatible client. Once the server is running, a Cursor or Claude Desktop configuration simply points to it via . The AI client then calls the defined tools as if they were native language features. Because all communication is mediated by MCP, developers retain full control over authentication and data flow, while benefiting from the expressive power of AI to interpret and act on complex test telemetry.

Unique advantages include a dedicated focus on Currents’ data model, ensuring that every tool maps directly to meaningful test artifacts. The server also handles authentication via environment variables, keeping secrets out of the codebase. By coupling Currents’ granular test insights with AI’s reasoning capabilities, this MCP server transforms raw CI output into intelligent, context‑aware guidance that accelerates debugging and improves software quality.