Skip to content

Releases: Santosh-Prasad-Verma/ClaudeGate

Release list

v1.1.0 — Universal Agent Gateway & Minimalist Console

v1.1.0 — Universal Agent Gateway & Minimalist Console Latest
Latest

Choose a tag to compare

View all tags
@Santosh-Prasad-Verma Santosh-Prasad-Verma released this 13 Sep 23:17
v1.1.0
9b392db
This commit was created on GitHub.com and signed with GitHub’s verified signature.
GPG key ID: B5690EEEBB952194
Verified
Learn about vigilant mode.

What's New in v1.1.0

Universal Agent Support

  • OpenAI Codex CLI Support: Added native wire compatibility for the /v1/responses API endpoint used by modern Codex CLI (v0.154.0+).
  • Universal OpenAI Endpoint: Added /v1/chat/completions for seamless compatibility with OpenCode, Aider, CommandCode, and Cursor.
  • Stream-Through Tool Execution: Zero-delay argument streaming to eliminate client stalls and prevent empty {} crashes.
  • Extended Thinking & Reasoning: Automatic translation of DeepSeek reasoning_content and Groq reasoning into native Anthropic thinking events, with automatic stripping of in-band <think> tags.

Minimalist SVG Live Console Dashboard

  • Redesigned /dashboard with a clean, high-contrast monochrome interface.
  • Strictly Zero Emojis: Crisp inline SVGs across the entire interface.
  • Strictly Zero Hover Animations: Removed CSS transitions, keyframes, and translateY transforms.
  • Real-Time Flight Recorder: In-memory ring buffer tracking all incoming requests, models, latency, and tool invocations.
  • Local Health & Ping Probe: Instant connectivity test with upstream latency measurements.

Production Resiliency & Hardening

  • Anti-ECONNRESET Heartbeat Pings: Compliant 15-second SSE keep-alive pings to keep agent connections alive during extended reasoning periods.
  • Context Overflow Guard: Automatic compaction of ancient intermediate tool outputs (>1000 characters) when nearing context limits.
  • 3-State Zero-Latency Circuit Breaker: Instant failover to secondary providers on upstream outages or 429 rate limits.
  • Multi-Host Tier Routing: Independent base URLs and API keys for Big, Mid, and Small model tiers.
  • OpenAI o1/o3/o4 Compatibility: Automatic mapping to max_completion_tokens and reasoning_effort.
  • Vision Fallback: Automatic textual fallback for image inputs sent to text-only code models.

Developer Tooling & Security

  • System & Model Doctor: Run python start_proxy.py --doctor to scan local models, check connectivity, and validate function-calling capability.
  • Security Remediation: Fixed CVEs across dependencies (setuptools, wheel, msgpack, pygments, logfire) and hardened schema sanitizers.
  • Thorough Documentation: Completely revamped README.md covering all agent setups, architecture, and configuration options.
Assets 2
Loading

v1.0.0 — Universal LLM Gateway & Bridge for Claude Code CLI

v1.0.0 — Universal LLM Gateway & Bridge for Claude Code CLI

Choose a tag to compare

View all tags
@Santosh-Prasad-Verma Santosh-Prasad-Verma released this 25 Aug 08:26
v1.0.0
9c44716
This commit was created on GitHub.com and signed with GitHub’s verified signature.
GPG key ID: B5690EEEBB952194
Verified
Learn about vigilant mode.

🎉 Welcome to ClaudeGate v1.0.0

ClaudeGate is a high-throughput, secure local API gateway bridging Anthropic's Messages API (/v1/messages) to any OpenAI-compatible provider. Power your Claude Code CLI, Cursor, and Anthropic SDK applications with OpenRouter, DeepSeek, Qwen, Kimi, or 100% private local models.

✨ Key Features in this Release:

  • ⚡ Zero-Crash SSE Streaming: Real-time token streaming translated into Anthropic SSE format without ASGI crashes or socket drops.
  • 🛠️ Full Bi-Directional Tool Calling: Seamlessly executes Claude Code terminal commands, file edits, and codebase navigation.
  • 🧹 <thinking> / CoT Sanitizer: Removes reasoning tokens and thinking blocks from conversation history to prevent multi-turn 400 Bad Request errors.
  • 🛡️ PII & Secret Redactor: Automatically intercepts and scrubs API keys and sensitive tokens before leaving your machine.
  • 🔄 Multi-Provider Failover: Graceful fallback between primary and backup upstream providers during rate limits or server timeouts.
  • 🎛️ 24+ Pre-Configured Presets: Instant configuration switching via --preset <name>.
  • 🐳 Docker Ready: Standalone container support with automated health-check endpoints.

🧠 Supported Providers & Models:

  • Cloud & Frontier: 0x Alpha, OpenRouter, DeepSeek (V4-Pro / V3 / R1), Alibaba Qwen (Qwen3.8 / Qwen 2.5 Coder), Moonshot Kimi (K3 / K2.7), MiniMax, Google Gemini, OpenAI, Mistral AI, Cohere, Perplexity.
  • Local & Offline: Ollama, LM Studio, vLLM, LocalAI (Zero data leaves your machine).

🚀 Quick Start:

git clone https://github.com/Santosh-Prasad-Verma/ClaudeGate.git
cd ClaudeGate
pip install -r requirements.txt

# Run interactive setup or use a preset:
python -m src.cli --setup
Loading