A tiny caching proxy layer for Matrix to allow a large number of clients to share a single access_token. Useful for deployments with extremely high read-only guest load.
Only a couple endpoints for reading room events and state are supported. The proxy is aware of Matrix's sync and events long-polling endpoints and batches all waiting client long-polls behind a single request to the Matrix backend.
PORT: Specify the port the http server will listen on.
CACHE_TTL_SECONDS: Duration to cache successful responses.
MATRIX_SERVER: Matrix server endpoint.
MATRIX_ACCESS_TOKEN: Access token for requests.