API Reference¶
What you'll learn here: which Python modules are meant for normal integrators, which ones are implementation details, and where to start reading the code through generated API docs.
How to use this section¶
This API reference is split into two parts on purpose.
- Public API is the small surface area that most readers should start with.
- Internal API covers the modules that make the adapter work, but are more likely to change as the implementation evolves.
If you are trying to embed the adapter, load config, or understand the top-level runtime entry points, stay in the public section first.
If you are contributing to the adapter itself, debugging proxy behavior, or tracing state and telemetry flows, the internal section is the right place.
Public API¶
These modules are the best first stop for normal users and integrators:
remote_mcp_adapter.server- app factory and top-level wiringremote_mcp_adapter.config.load- YAML loading and environment interpolationremote_mcp_adapter.config.schemas.root- the top-level validated config contract
Go to Public API.
Internal API¶
These modules are more implementation-oriented:
remote_mcp_adapter.proxy.factory- per-server proxy construction and session-pinned upstream clientsremote_mcp_adapter.proxy.hooks- adapter tool/resource override wiringremote_mcp_adapter.telemetry.manager- async telemetry collection and exportremote_mcp_adapter.core.storage.store- session, upload, artifact, cleanup, and quota state handling
Go to Internal API.
Reading order¶
If you are new to the codebase, this order is the least confusing:
Next steps¶
- Previous topic: Troubleshooting - common failures and practical fixes.
- Next: Public API - generated reference for the main entry points.