taskai

A command-line task manager with AI features.

📖 Full documentation: https://taskai.readthedocs.io

pip install taskai-cli

Setup

task setup

Commands

task                              interactive mode
task help                         show every command
task examples                     worked command sequences for common workflows

task show all                     show everything
task show <id|name>               show a specific item

task create <name>                create a top-level item
task add <parent> <name>          add a child item
task update <id> --field value    update any field
task rename <id> <new name>       rename an item
task move <item> <new parent>     reparent an item
task complete <id>                mark complete (recursive)
task delete <id|name>             delete by id or name
task clear                        delete all completed items
task clear <parent>               delete completed under a parent

task comment <id> <text>          add a comment
task link <parent> <item>         soft-link an item under a second parent
task unlink <parent> <item>       remove a soft-link
task reorder <id1> before|after <id2>

task status <id> <text>           set a status string
task pomo <on_mins> <off_mins>    pomodoro timer

task ai <prompt>                  natural language → commands
  --context <path[,path...]>      fold file contents into the prompt
  --reasoning <level>             model thinking effort: minimal|low|medium|high|disable|none
task ai headstart <id>            AI suggests next step, saved as comment

task browser                      launch the web UI (see below)

task config show                  list current config
task config set <key> <value>     set a config value
task config get <key>             get a config value
task config pop <key>             remove a config value

task repair                       repair the database
task nuke                         delete all data (asks to confirm)

Item fields

Pass any of these as --field value to create, add, or update:

Config

AI features need a model. task setup walks you through choosing a provider and model, then tells you which environment variable(s) to set:

$ task setup
Beginning setup
  1. openai
  2. anthropic
  3. gemini
  4. vertex_ai
  5. groq
  ...
  20. ollama

Select a provider (number or name): 3
  1. gemini-3.7-flash
  2. gemini-3.1-pro-preview
  3. gemini-3.6-flash
  4. gemini-3.5-flash-lite
  5. gemini-2.5-pro

Select a model for 'gemini' (number or name): 1
Storing: gemini/gemini-3.7-flash
Make sure these environment variables are set:
  GEMINI_API_KEY
Setup complete! Use 'task config set|get|list' to interact with your configuration options

Either prompt also accepts free-typed text, so you can pick a model newer than the built-in menu.

Or set it directly — a provider/model string, using litellm's naming:

task config set AI_MODEL <provider>/<model>   # e.g. gemini/gemini-3.7-flash

Web UI

task browser

Launches a local web server with a canvas view of your item tree:

• pan / zoom the graph; completed items are tinted green, and soft-links (task link) show as dashed "shadow" copies under their host item
• click a node to open an edit panel — each field writes back immediately, there's no save button
• console at the bottom for typing raw task commands against the same data

Examples

task create "Launch website"

# add subtasks (by name or id)
task add "Launch website" "Write copy"
task add "Launch website" "Design mockups" --priority 1
task add "Launch website" "Set up hosting" --due_by 08-15-2026

# check your work
task show "Launch website"

# ask AI what to do first on a specific task
task ai headstart 3

# mark something done and clean up
task complete 2
task clear "Launch website"

Or drive it entirely from a prompt:

task ai "add a task called 'Reply to Alex' under Launch website, and mark the hosting task as done"

Run task examples for more worked sequences.