todo command and --todo flag help you manage task lists and automatically generate todos from agent responses.
Install requirement:
praisonai todo requires pip install praisonai (the full wrapper). On a standalone pip install praisonai-code install every todo subcommand exits 1 with todo requires the full wrapper. Install the full wrapper: pip install praisonai — a single-line hint, no Rich traceback (PR #2854).Quick Start


Commands
Add Todo
List Todos
Complete Todo
Delete Todo
Clear Todos
Help
Using —todo Flag
Generate Todos from Task
Combine with Planning
Use Cases
Project Planning
Sprint Planning
Learning Path
Bug Triage
Todo Storage
Todos are stored in the home-level workspace~/.praison/todos.json:
The todo workspace (
~/.praison/) is a home-level location and was not changed by PraisonAI PR #4946 — it does not trigger the project-marker path drift. However, project-level CLI state — thinking, compaction, output-style, file-history, the queue DB, and knowledge vector stores — now lives under ./.praisonai/ (canonical). See Project data directory.Priority Levels
Best Practices
Regular Review
Use
praisonai todo list regularly to track progressClear Completed
Periodically clear completed todos to keep list manageable
Combine with Planning
Use
--planning --todo for detailed task breakdownsProject-Specific
Consider using different workspaces for different projects
Workflow Example
Via MCP
Thepraisonai-mcp host exposes four MCP tools — praisonai.todo.list, praisonai.todo.add, praisonai.todo.complete, praisonai.todo.delete — that read and write the runtime todo store (<workspace>/todos.json, else ~/.praisonai/todos.json). A todo added from Claude Desktop is visible to an agent run on the same machine, and vice versa.
See MCP Todo Tools for the tool table, record shape, and store resolution.

