Skip to main content
pflow is a workflow compiler for AI agents. Your agent reasons through a task once, pflow compiles it into a workflow, and that workflow runs instantly forever after - no re-planning, no repeated costs.

​
The problem

AI agents are powerful but inefficient for repetitive tasks:
  • Re-reasoning: Every time you ask an agent to do the same task, it reasons through it from scratch
  • Cost: Each reasoning pass costs tokens - the same tokens, for the same logic, every time
  • Inconsistency: Agents might take different paths or skip steps on different runs
  • Context bloat: Loading tool schemas (especially MCP servers) consumes tokens before any work begins

​
How pflow helps

pflow separates planning from execution:
  1. Your agent plans once - figures out what nodes to use and how to connect them
  2. pflow compiles the workflow - saves it as a reusable pipeline
  3. Execution is instant - run the same workflow with different inputs, zero reasoning cost
Nodes are individual tools (call an LLM or API, read a file, or any MCP tool). Workflows chain them together. Over time, your library grows - complex tasks become instant commands.

​
Two ways to run workflows

Agent-driven: Your AI agent discovers, creates, and runs workflows through pflow. This is the primary flow - your agent handles automation tasks using workflows it has built. Direct CLI: Saved workflows are CLI commands. Run them directly in scripts, CI/CD pipelines, or cron jobs - no agent needed. 
# Agent builds this once, then anyone can run it
pflow analyze-logs input=./logs/api.log format=json

​
Get started

Quickstart

Install pflow and connect your AI tool

Integrations

Set up Claude Code, Cursor, VS Code, and more

Adding MCP servers

Expand pflow with external tools

CLI reference

All commands documented