Higress Plugin Marketplace

Unified AI model proxy supporting multiple LLM providers with protocol conversion and load balancing

LLM response caching with semantic and string-matching support to reduce API costs

Token-based rate limiting for precise AI cost and concurrency control

Manage AI usage quotas for consumers with multi-dimensional quota policies

Intercept or replace sensitive data in requests/responses to protect privacy

Automatically add system prompts or prefixes/suffixes to enhance model output

Use predefined prompt templates to simplify AI application development

Integrate with Alibaba Cloud content security for AI input/output safety detection

Collect AI call metrics including token usage, latency and multi-dimensional monitoring

Support AI Agent mode with tool calling and multi-turn conversation orchestration

Provide web search capabilities for AI models to enhance real-time information retrieval

Force AI to return structured JSON format for easier downstream parsing

Automatically manage multi-turn conversation history with context window control

Identify user intent for intelligent routing and traffic splitting

Retrieval-Augmented Generation to enhance AI answer accuracy with knowledge bases

Flexibly transform AI request/response formats to adapt different model protocols

Parse client IP geolocation for region-based access control

JWT-based authentication with multi-consumer configuration and fine-grained access control

Simple API Key-based authentication supporting multiple key delivery methods

HTTP Basic authentication with username and password verification

HMAC signature-based authentication for enhanced security

OAuth 2.0 protocol support for third-party authorization

OpenID Connect authentication with major identity provider integration

Forward authentication requests to external services for custom auth logic

Open Policy Agent-based policy control for flexible access decisions

APISIX-compatible HMAC authentication format for easy migration

ModSecurity-based WAF with OWASP CRS rules support

Configure CORS response headers to control cross-origin resource access

IP-based blacklist/whitelist control to restrict or allow specific IP access

Detect and block bot traffic to protect services from crawler attacks

Block specific requests based on rules with URL, Header and multi-dimensional matching

Prevent request replay attacks to protect API security

Key-based distributed cluster rate limiting with Redis storage

Key-based local rate limiting without external dependencies

Add tags to traffic for canary releases and traffic grouping

Validate request format and parameters to ensure compliance with expected specifications

Flexibly transform request/response content with Header and Body modifications

Convert RESTful requests to GraphQL queries

Customize response content for mock testing or static responses

Frontend resource canary release with multi-version parallel deployment

Set browser cache response headers to optimize resource loading performance