https://praxis.fast/ai/dev/filters/Recent content in Filters on PraxisHugoenhttps://praxis.fast/ai/dev/filters/a2a/Mon, 01 Jan 0001 00:00:00 +0000https://praxis.fast/ai/dev/filters/a2a/Extracts A2A protocol metadata from JSON-RPC request bodies and promotes method, family, task ID, streaming detection, and version to request headers, filter results, and durable metadata for routing.https://praxis.fast/ai/dev/filters/ai_guardrails/Mon, 01 Jan 0001 00:00:00 +0000https://praxis.fast/ai/dev/filters/ai_guardrails/Calls an external AI guardrail provider to evaluate request and response bodies. The provider determines whether content should be passed, blocked, or redacted.https://praxis.fast/ai/dev/filters/ai_operation/Mon, 01 Jan 0001 00:00:00 +0000https://praxis.fast/ai/dev/filters/ai_operation/Classifies supported AI operations from the request head.https://praxis.fast/ai/dev/filters/anthropic_messages_protocol/Mon, 01 Jan 0001 00:00:00 +0000https://praxis.fast/ai/dev/filters/anthropic_messages_protocol/Normalizes Anthropic Messages protocol headers for native backends.https://praxis.fast/ai/dev/filters/anthropic_messages_request/Mon, 01 Jan 0001 00:00:00 +0000https://praxis.fast/ai/dev/filters/anthropic_messages_request/Processes an Anthropic create-message body once and publishes its facts.https://praxis.fast/ai/dev/filters/anthropic_messages_to_chat_completions/Mon, 01 Jan 0001 00:00:00 +0000https://praxis.fast/ai/dev/filters/anthropic_messages_to_chat_completions/Transforms Anthropic Messages API requests to Chat Completions-compatible request bodies and transforms compatible responses back.https://praxis.fast/ai/dev/filters/anthropic_messages_to_chat_completions_stream/Mon, 01 Jan 0001 00:00:00 +0000https://praxis.fast/ai/dev/filters/anthropic_messages_to_chat_completions_stream/Transforms streaming SSE responses between the Chat Completions and Anthropic Messages formats, processing each chunk as it arrives.https://praxis.fast/ai/dev/filters/anthropic_messages_to_vertexai_anthropic/Mon, 01 Jan 0001 00:00:00 +0000https://praxis.fast/ai/dev/filters/anthropic_messages_to_vertexai_anthropic/Translates Anthropic Messages requests to Vertex AI rawPredict and Vertex responses back to the Anthropic dialect.https://praxis.fast/ai/dev/filters/anthropic_web_search/Mon, 01 Jan 0001 00:00:00 +0000https://praxis.fast/ai/dev/filters/anthropic_web_search/Executes server-owned WebSearch tool calls in an Anthropic Messages loop.https://praxis.fast/ai/dev/filters/aws_sigv4_sign/Mon, 01 Jan 0001 00:00:00 +0000https://praxis.fast/ai/dev/filters/aws_sigv4_sign/Signs outbound requests to AWS services using SigV4.https://praxis.fast/ai/dev/filters/azure_ad/Mon, 01 Jan 0001 00:00:00 +0000https://praxis.fast/ai/dev/filters/azure_ad/Injects an Azure AD (Entra ID) bearer token into outbound requests.https://praxis.fast/ai/dev/filters/callout_credentials/Mon, 01 Jan 0001 00:00:00 +0000https://praxis.fast/ai/dev/filters/callout_credentials/Establishing filter that captures typed per-user callout secrets from ingress headers.https://praxis.fast/ai/dev/filters/credential_inject/Mon, 01 Jan 0001 00:00:00 +0000https://praxis.fast/ai/dev/filters/credential_inject/Replaces caller credentials with the upstream credential selected by intelligent_route or provider_route.https://praxis.fast/ai/dev/filters/extensions/Mon, 01 Jan 0001 00:00:00 +0000https://praxis.fast/ai/dev/filters/extensions/Extend Praxis AI with custom filters using the Praxis core filter interfaces and registration model.https://praxis.fast/ai/dev/filters/external_metering/Mon, 01 Jan 0001 00:00:00 +0000https://praxis.fast/ai/dev/filters/external_metering/Integrates with an external metering service for pre-request balance checks and post-response token usage reporting.https://praxis.fast/ai/dev/filters/reference/Mon, 01 Jan 0001 00:00:00 +0000https://praxis.fast/ai/dev/filters/reference/AI filters provided by Praxis AI. For base proxy filters (router, load balancer, headers, CORS, etc.), see the Praxis core filter reference.https://praxis.fast/ai/dev/filters/gcp_adc/Mon, 01 Jan 0001 00:00:00 +0000https://praxis.fast/ai/dev/filters/gcp_adc/Injects a GCP OAuth2 access token into outbound requests.https://praxis.fast/ai/dev/filters/http_callout/Mon, 01 Jan 0001 00:00:00 +0000https://praxis.fast/ai/dev/filters/http_callout/Calls an external HTTP service during request processing and feeds its response into branch-chain evaluation.https://praxis.fast/ai/dev/filters/identity_header_guard/Mon, 01 Jan 0001 00:00:00 +0000https://praxis.fast/ai/dev/filters/identity_header_guard/Captures request headers matching a configured prefix into filter_metadata and removes them from the upstream request.https://praxis.fast/ai/dev/filters/inflight_tracker/Mon, 01 Jan 0001 00:00:00 +0000https://praxis.fast/ai/dev/filters/inflight_tracker/Tracks live request counts and reserved tokens per model.https://praxis.fast/ai/dev/filters/intelligent_route/Mon, 01 Jan 0001 00:00:00 +0000https://praxis.fast/ai/dev/filters/intelligent_route/Selects an upstream cluster from a site/capability descriptor by matching either an inference model name or MCP tool name.https://praxis.fast/ai/dev/filters/llmisvc_model_provider_resolver/Mon, 01 Jan 0001 00:00:00 +0000https://praxis.fast/ai/dev/filters/llmisvc_model_provider_resolver/Rewrites publisher-ID body model values to the short model name for LLMISvc / KServe routing; the routing header is left unchanged.https://praxis.fast/ai/dev/filters/mcp/Mon, 01 Jan 0001 00:00:00 +0000https://praxis.fast/ai/dev/filters/mcp/Extracts MCP JSON-RPC method, tool, resource, prompt, version, and session metadata into filter-visible request fields.https://praxis.fast/ai/dev/filters/model_to_header/Mon, 01 Jan 0001 00:00:00 +0000https://praxis.fast/ai/dev/filters/model_to_header/Promotes the JSON “model” field from the request body to a request header.https://praxis.fast/ai/dev/filters/model_to_provider/Mon, 01 Jan 0001 00:00:00 +0000https://praxis.fast/ai/dev/filters/model_to_provider/Maps stable client-facing model IDs to existing upstream clusters and provider-specific target models.https://praxis.fast/ai/dev/filters/openai_agentic_loop/Mon, 01 Jan 0001 00:00:00 +0000https://praxis.fast/ai/dev/filters/openai_agentic_loop/Agentic loop controller for the Responses API pipeline.https://praxis.fast/ai/dev/filters/openai_chat_completions_request/Mon, 01 Jan 0001 00:00:00 +0000https://praxis.fast/ai/dev/filters/openai_chat_completions_request/Publishes body facts for native Chat Completions create requests to durable metadata and filter results.https://praxis.fast/ai/dev/filters/openai_chat_completions_to_anthropic/Mon, 01 Jan 0001 00:00:00 +0000https://praxis.fast/ai/dev/filters/openai_chat_completions_to_anthropic/Translates OpenAI Chat Completions requests only when the selected ExternalModel candidate uses Anthropic Messages; OpenAI candidates pass through with their configured path and model.https://praxis.fast/ai/dev/filters/openai_chat_completions_to_anthropic_stream/Mon, 01 Jan 0001 00:00:00 +0000https://praxis.fast/ai/dev/filters/openai_chat_completions_to_anthropic_stream/Converts the selected Anthropic provider’s event stream incrementally. OpenAI-compatible candidates are left byte-for-byte unchanged.https://praxis.fast/ai/dev/filters/openai_chat_completions_to_azureai_chat_completions/Mon, 01 Jan 0001 00:00:00 +0000https://praxis.fast/ai/dev/filters/openai_chat_completions_to_azureai_chat_completions/Transforms requests targeting Azure OpenAI deployments into standard Chat Completions-compatible form and normalizes responses back.https://praxis.fast/ai/dev/filters/openai_chat_completions_to_vertexai_gemini/Mon, 01 Jan 0001 00:00:00 +0000https://praxis.fast/ai/dev/filters/openai_chat_completions_to_vertexai_gemini/Transforms OpenAI Chat Completions requests into Vertex AI Gemini generateContent format and translates responses back.https://praxis.fast/ai/dev/filters/openai_client_tool_compat/Mon, 01 Jan 0001 00:00:00 +0000https://praxis.fast/ai/dev/filters/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.https://praxis.fast/ai/dev/filters/openai_conversations/Mon, 01 Jan 0001 00:00:00 +0000https://praxis.fast/ai/dev/filters/openai_conversations/Serve OpenAI Conversations API endpoints locally through the configured response store.https://praxis.fast/ai/dev/filters/openai_doc_extract/Mon, 01 Jan 0001 00:00:00 +0000https://praxis.fast/ai/dev/filters/openai_doc_extract/Converts input_file content parts to input_text for backends that do not support input_file natively (e.g. vLLM, llm-d).https://praxis.fast/ai/dev/filters/openai_file_resolve/Mon, 01 Jan 0001 00:00:00 +0000https://praxis.fast/ai/dev/filters/openai_file_resolve/Resolves file_id and file_url references in Responses API input by fetching content from a Files API or remote URL via ApiClient and inlining the base64-encoded content in the provider-native field.https://praxis.fast/ai/dev/filters/openai_file_search_dispatch/Mon, 01 Jan 0001 00:00:00 +0000https://praxis.fast/ai/dev/filters/openai_file_search_dispatch/Dispatches the loop owner’s pending file-search assignments against a vector store API compatible backend.https://praxis.fast/ai/dev/filters/openai_mcp_dispatch/Mon, 01 Jan 0001 00:00:00 +0000https://praxis.fast/ai/dev/filters/openai_mcp_dispatch/Executes MCP tool calls against upstream MCP servers within the Responses API agentic loop.https://praxis.fast/ai/dev/filters/openai_mcp_tool_resolve/Mon, 01 Jan 0001 00:00:00 +0000https://praxis.fast/ai/dev/filters/openai_mcp_tool_resolve/Resolves MCP tool entries from the Responses API tools array into concrete tool definitions by calling tools/list on each upstream MCP server.https://praxis.fast/ai/dev/filters/openai_responses_compact/Mon, 01 Jan 0001 00:00:00 +0000https://praxis.fast/ai/dev/filters/openai_responses_compact/Summarizes conversation history when the token count exceeds a configured threshold.https://praxis.fast/ai/dev/filters/openai_responses_model_rewrite/Mon, 01 Jan 0001 00:00:00 +0000https://praxis.fast/ai/dev/filters/openai_responses_model_rewrite/Rewrites the model field in Responses and Chat Completions request bodies.https://praxis.fast/ai/dev/filters/openai_responses_proxy/Mon, 01 Jan 0001 00:00:00 +0000https://praxis.fast/ai/dev/filters/openai_responses_proxy/Rebuilds the request body from ResponsesState when present.https://praxis.fast/ai/dev/filters/openai_responses_rehydrate/Mon, 01 Jan 0001 00:00:00 +0000https://praxis.fast/ai/dev/filters/openai_responses_rehydrate/Validates previous_response_id by fetching the stored response, confirming its status is “completed”, and populating ResponsesState with the full conversation history (stored turns + current input).https://praxis.fast/ai/dev/filters/openai_responses_request/Mon, 01 Jan 0001 00:00:00 +0000https://praxis.fast/ai/dev/filters/openai_responses_request/Processes a Responses request body once and initializes state.https://praxis.fast/ai/dev/filters/openai_responses_store/Mon, 01 Jan 0001 00:00:00 +0000https://praxis.fast/ai/dev/filters/openai_responses_store/Persists Responses API responses to the configured response store backend.https://praxis.fast/ai/dev/filters/openai_responses_to_chat_completions/Mon, 01 Jan 0001 00:00:00 +0000https://praxis.fast/ai/dev/filters/openai_responses_to_chat_completions/Translates canonical Responses create requests for a Chat Completions backend.https://praxis.fast/ai/dev/filters/openai_stream_events/Mon, 01 Jan 0001 00:00:00 +0000https://praxis.fast/ai/dev/filters/openai_stream_events/Composes the current IRR execution into one logical Responses stream.https://praxis.fast/ai/dev/filters/openai_tool_parse/Mon, 01 Jan 0001 00:00:00 +0000https://praxis.fast/ai/dev/filters/openai_tool_parse/Parses tool definitions and tool_choice from Responses API request bodies and promotes routing facts to metadata and filter results without mutating the body.https://praxis.fast/ai/dev/filters/openai_web_search_dispatch/Mon, 01 Jan 0001 00:00:00 +0000https://praxis.fast/ai/dev/filters/openai_web_search_dispatch/Web search filter for model-driven web_search_call dispatch.https://praxis.fast/ai/dev/filters/project_state_owner_headers/Mon, 01 Jan 0001 00:00:00 +0000https://praxis.fast/ai/dev/filters/project_state_owner_headers/Projects a normalized [StateOwner] into destination-specific HTTP headers.https://praxis.fast/ai/dev/filters/prompt_enrich/Mon, 01 Jan 0001 00:00:00 +0000https://praxis.fast/ai/dev/filters/prompt_enrich/Injects statically configured messages into the messages array of OpenAI-compatible chat completion request bodies.https://praxis.fast/ai/dev/filters/provider_route/Mon, 01 Jan 0001 00:00:00 +0000https://praxis.fast/ai/dev/filters/provider_route/Exact provider-local mapping from an authenticated intelligent routing selection to a private backend cluster.https://praxis.fast/ai/dev/filters/state_owner/Mon, 01 Jan 0001 00:00:00 +0000https://praxis.fast/ai/dev/filters/state_owner/Establishes a normalized [StateOwner] from trusted identity sources.https://praxis.fast/ai/dev/filters/stream_usage_inject/Mon, 01 Jan 0001 00:00:00 +0000https://praxis.fast/ai/dev/filters/stream_usage_inject/Injects stream_options.include_usage = true into streaming OpenAI chat-completions requests so the upstream response contains token usage.https://praxis.fast/ai/dev/filters/time_to_first_token/Mon, 01 Jan 0001 00:00:00 +0000https://praxis.fast/ai/dev/filters/time_to_first_token/Measures time-to-first-token for streaming AI responses.https://praxis.fast/ai/dev/filters/token_ceiling/Mon, 01 Jan 0001 00:00:00 +0000https://praxis.fast/ai/dev/filters/token_ceiling/Rejects provider-bound requests whose estimated serialized request body or requested output exceeds configured limits.https://praxis.fast/ai/dev/filters/token_count/Mon, 01 Jan 0001 00:00:00 +0000https://praxis.fast/ai/dev/filters/token_count/Extracts token usage from AI inference responses and writes unified counts to [filter_metadata].https://praxis.fast/ai/dev/filters/token_rate_limit/Mon, 01 Jan 0001 00:00:00 +0000https://praxis.fast/ai/dev/filters/token_rate_limit/Token-denominated rate limiter: reserves an estimated cost at admission, reconciles against actual usage after the response completes.https://praxis.fast/ai/dev/filters/token_usage_headers/Mon, 01 Jan 0001 00:00:00 +0000https://praxis.fast/ai/dev/filters/token_usage_headers/Injects Praxis-Token-Input, Praxis-Token-Output, and Praxis-Token-Total headers into downstream responses when token usage data is present in [filter_metadata].