Configuration#
PyFluent-MCP is configured primarily through FLUIDS_MCP_* environment variables.
You can set these in your shell, MCP client configuration, or a .env file loaded by
your client.
General settings#
Variable |
Effect |
|---|---|
|
External JSON file for overriding the bundled settings schema |
|
Server log level (default |
|
|
|
Cap on deterministic tool-loop iterations (default |
|
|
Execution and retrieval behavior#
PyFluent-MCP focuses on deterministic Fluent tooling and MCP transport exposure.
The repository provides:
validated Fluent execution
schema retrieval and grounding
backend abstraction
deterministic tool execution
MCP transport support
Transport security#
TLS certificate verification is enabled by default for outbound retrieval and HTTP operations.
Variable |
Purpose |
|---|---|
|
Path to a PEM CA bundle (also reads |
|
Enable or disable TLS verification (default enabled) |
|
Timeout for outbound HTTP operations |
Server command-line tool options#
The ansys-fluent-mcp console script accepts:
You can also launch the server as a module:
python -m ansys.fluent.mcp
Flag |
Description |
|---|---|
|
MCP transport (default |
|
HTTP bind address (when |
|
Default backend kind until |
|
Log level (default |
Next steps#
For passing environment variables through MCP clients, see IDE and client configuration.