Skip to content

v0.1.0 — Initial public release

Choose a tag to compare

View all tags
@linxule linxule released this 01 Mar 14:43
· 15 commits to main since this release
v0.1.0
bd03f5f

First public release of Vox MCP.

Features

  • 8 providers: Google Gemini, OpenAI, Anthropic, xAI (Grok), DeepSeek, Moonshot (Kimi), OpenRouter, and custom OpenAI-compatible endpoints (Ollama, vLLM, LM Studio)
  • 3 tools: chat (multi-model gateway with file/image context), listmodels (model discovery), dump_threads (thread export as JSON or Markdown)
  • Conversation memory: in-memory threads with continuation_id for multi-turn exchanges across any provider
  • Thread persistence: append-only JSONL with cold-reload from disk
  • Thinking mode support: per-model constraint architecture (token budgets, effort levels, always-on reasoning)
  • Temperature constraints: fixed, range, and discrete handling per model
  • Model restrictions: per-provider allowlists via environment variables
  • Auto mode: agent-driven model selection when DEFAULT_MODEL=auto

Design

Pure API passthrough — no system prompt injection, no response modification. Conversation memory is the sole value-add.

Install

git clone https://github.com/linxule/vox-mcp.git
cd vox-mcp
cp .env.example .env  # add at least one API key
uv sync
uv run python server.py

Also available on PyPI and the MCP Registry.

Attribution

Derived from pal-mcp-server by Beehive Innovations. Apache 2.0.

Assets 2
Loading