Terminal UI, Streaming & Latency Metric

[0.22.0] - 2026-02-14

Added

• Basic Terminal UI via agentify shell

Latency Metric

• Latency tracking for model invocations
• agentify runtime list displays latency metrics
• agentify run --showstats for token usage, cost, and now latency

Streaming

• Added streaming support via agentify run --stream
• Supported providers:
  • OpenAI
  • Mistral
  • DeepSeek
  • X AI
  • GitHub Models
  • Ollama (Cloud and Local)
• Unified provider interface:

  run_provider(model_id, user_prompt, stream=True)```