openai_client_tool_compat
Lowers rich client-owned tool declarations to private function tools for a function-only Responses backend and restores the typed items on the way back.
On this page
Lowers rich client-owned tool declarations to private function tools for a function-only Responses backend and restores the typed items on the way back.
Configuration Notes
Enablement is by placement in the pipeline routed to a function-only Responses backend, not by a “target backend” flag. Config fields bound behaviour (byte and tool-count caps) only, mirroring the other Responses rewriter filters.
Configuration
| Field | Type | Required | Description |
|---|---|---|---|
max_rewritten_body_bytes | integer | no | Maximum size in bytes of the request body this filter produces after lowering rich client tool declarations to function entries, and of the response body it produces after restoring typed client-owned items. Raw request body size is governed by the pipeline’s body_limits, not this field. Lowering absorbs provider fields into function descriptions, so the rewritten body can grow larger than the raw input. |
max_client_tools | integer | no | Maximum number of client-owned tool declarations lowered per request. |
Examples
Example 1
filter: openai_client_tool_compat
Example 2
filter: openai_client_tool_compat
max_rewritten_body_bytes: 67108864
max_client_tools: 512