proxy

MorrowCache is a local OpenAI-compatible proxy that caches AI model responses based on intent similarity to reduce latency and token costs, using an adjudicator to check for cached answers.

The tweet shares a repository with a transparent guide for setting up a stable proxy strategy for AI traffic, using tools like Hysteria2 and WARP without overpromising safety.

A GitHub repository has been created that redirects Claude Code traffic to free providers like DeepSeek and Kimi, enabling users to run it for free, with over 20,000 developers already using it.

This article describes a method to train LoRA adapters using AsyncGRPOTrainer and sync them via Storage Buckets across separate Hugging Face Jobs, eliminating the need for NCCL communication.

Routeup provides stable, browser-trusted HTTPS names for local development apps, with opt-in public tunnels that require no port forwarding or router configuration.

The author is seeking feedback on a self-serve proxy service designed to enhance SEO and AI agent readability by serving clean HTML, dynamically adding meta tags, and offering analytics.

The article discusses the need for tool gateways to secure AI agents' API access by limiting unpredictable behavior and seeks recommendations for products or libraries that provide such functionality.

FAAAH is a dependency-free, OpenAI-compatible proxy that converts API requests into text files and uses AI coding agents like Claude Code as the backend, enabling reuse of existing subscriptions instead of paying for cloud LLM APIs.

free-claude-code is an open-source proxy tool that runs Claude Code, Codex, and Pi through its own provider without requiring an API key. It supports 31 cloud and local models, as well as multiple access methods such as terminal, IDE, and mobile. It has currently gained 43.8k stars on GitHub.

Discusses why, among iOS proxy tools, the paid and closed-source Shadowrocket is more popular than the free and open-source Clash Mi.

Tailscale on jailbroken Kindles has been updated with proxy and TUN modes, allowing apps like KOReader to reach other Tailscale devices, and enabling Tailscale SSH by default.

opencodex is a lightweight local proxy that lets you use any LLM (Claude, Gemini, Grok, DeepSeek, etc.) with OpenAI Codex and Claude Code, supporting streaming, tool calls, and reasoning tokens.

Explains how to secure local AI agent tool execution in OpenClaw by using Loopers proxy and OPA/Rego policies to intercept and validate MCP tool calls before they execute on the host machine.

Plano is an open-source proxy that sits between AI agents and LLM providers to cut costs through intelligent routing, guardrail filtering, and cost-aware selection, all configured via a single YAML file without modifying agent code.

Explains why model routing in agent tasks may not save costs due to cache warmup, and describes a production solution with model affinity and the open-source proxy Plano to achieve actual savings.

pxpipe is a local open-source proxy tool that reduces Claude Code bills by approximately 70% by rendering large amounts of text (such as system prompts, code, and logs) into PNG images and feeding them through the large model's vision channel. It exploits the fact that images are billed by pixel rather than by word count. The tool is perfectly compatible with the Fable 5 model, operates with clever automation, but uses lossy compression and is unsuitable for sensitive data.

VisionBridge is an open-source proxy that gives text-only LLMs vision capabilities by letting a reasoning model query a separate vision model for image inspection, OCR, and more.

A developer built an open-source proxy (KU-Gateway) that drops stale context from vector database retrievals before LLM synthesis, cutting token burn by ~50% and preventing stale-data hallucinations. The tool is now opening for a 14-day stress test/hackathon.

Matt Pocock discovers it's trivial to build a proxy for reading raw system prompts sent to Anthropic from Claude Code, and he's now working to remove bloat in the system prompt.

A security expert discusses the challenge of securing identity tokens for LLM agents that access sensitive resources, and proposes a proxy-based approach to bind tokens to specific environments to prevent credential theft.