pawan 0.4.9

Pawan (पवन) — CLI coding agent binary
pawan-0.4.9 is not a library.

pawan v0.4.8

Pawan (पवन) — CLI coding agent with pluggable LLM backends, 34 tools, and cross-session memory.

What's New in v0.4.8

  • Shell-like command history navigation — Up/down arrow keys navigate through command history, matching bash/zsh/Claude Code/OMP behavior
  • History persistence — Commands saved to history (excluding slash commands), accessible across sessions

What's New in v0.4.7

  • Redesigned keyboard shortcuts — Ctrl+C always clears input, Ctrl+Q quits for cleaner separation of concerns

What's New in v0.4.6

  • Redesigned Ctrl+C behavior — Clears input when non-empty, quits when empty (matching bash/zsh pattern)

What's New in v0.4.1

  • Session tags UI — Visual green tags in status bar, manage via /tag command
  • Fuzzy session search — Fuzzy matching with [FUZZY] indicator in session browser
  • NVIDIA NIM catalog/models command to browse available NIM models
  • Enhanced /diff--cached flag support and colorized diff output
  • Improved /load and /resume — Opens session browser when called without arguments
  • Enhanced scrolling — PageUp/PageDown/Home/End keys and mouse wheel support in all popups (command palette, slash menu, session browser)
  • Secure credential storage — API keys stored in OS-native credential store (Keychain/Credential Manager/libsecret)
  • Test stability — All 722 library tests and 59 integration tests now pass consistently

Install

cargo install pawan

Features

  • Multi-model — NVIDIA NIM, Ollama, OpenAI, MLX backends
  • 34 tools in 3 tiers (Core: file ops, Standard: git/search, Extended: web/MCP)
  • Ratatui TUI with interleaved content blocks (tool calls inline with text)
  • Model Selector — Interactive model selection with search and filtering
  • Session Browser — Browse, load, and manage saved sessions
  • Auto-save — Automatic session saving at configurable intervals
  • MCP client via rmcp — connect to any MCP server
  • Eruka memory — cross-session context via Eruka integration
  • Thinking modes — per-model dispatch (Qwen, Gemma, Mistral, DeepSeek)

Usage

# Interactive mode
pawan

# With a specific model
pawan --model qwen/qwen3.5-122b-a10b

# Execute a task
pawan -e "fix the failing test in src/lib.rs"

# Use new slash commands
pawan  # then use /sessions, /save, /load, /resume, /new

License

MIT