signalsift

galacticsurfer/signalsift

★ 0 stars Python AI/LLM Updated 2d ago

Local-first CloudWatch observability MCP server: deterministic log reduction + local LLM analysis, compact incident reports for Claude/Cursor

View on GitHub → 🔍 Audit Wallet Slippage →

Quick Install

Copy the config for your editor. Some servers may need additional setup — check the README.

Add to claude_desktop_config.json:

{
  "mcpServers": {
    "signalsift": {
      "command": "uvx",
      "args": [
        "signalsift"
      ]
    }
  }
}

Or install with pip: pip install signalsift

README Excerpt

Local-first observability intelligence MCP server. SignalSift connects to AWS CloudWatch Logs, aggressively reduces noisy log data with deterministic Python processing, analyzes the reduced evidence with a small **local** LLM (Ollama on Apple Silicon), and returns a compact, evidence-validated incident report to an MCP client such as Claude Code or

Tools (17)

ALLOWED_LOG_GROUPSCACHE_PATHCACHE_TTL_SECONDSDEBUGLLM_MODELLLM_TIMEOUT_SECONDSMAX_LLM_INPUT_CHARSMAX_MCP_RESPONSE_CHARSMAX_QUERY_RESULTSMAX_TIME_RANGE_MINUTESOLLAMA_URLSIGNALSIFT_analyze_incidentcompare_windowslist_log_groupssearch_errorstrace_request

signalsift

Quick Install

README Excerpt

Tools (17)

Related AI/LLM Servers