8 tools. 6 can modify or destroy data without limits.
6 write tools that can modify data. Rate limits recommended.
Last updated:
Execute tools (codestral_fim, mistral_chat, mistral_ocr) trigger processes with side effects. Builds, notifications, workflows — all fired without throttling.
voxtral_transcribe:
rules:
- rate_limit: 60/minute Controls API costs and prevents retry loops from exhausting upstream rate limits.
The Mistral server is primarily read-only with 2 read tools. While it cannot modify data, an agent in a retry loop can make thousands of API calls per minute, exhausting rate limits and running up costs. Rate limiting is still recommended.
8 tools across 2 categories: Execute, Read. 2 are read-only. 6 can modify, create, or delete data.
One line change. Instead of running the Mistral server directly, prefix it with Intercept: intercept -c io-github-swih-mistral-mcp.yaml -- npx -y @mistral-mcp. Download a pre-built policy from policylayer.com/policies/io-github-swih-mistral-mcp and adjust the limits to match your use case.
Starter policies available for each. Same risk classification, same one-command setup.
Deterministic policy on every MCP tool call. Per-identity grants. Full audit log.