Pi Workflows Control Flow

Control flow inside Pi: TypeScript graphs with y/n branches, implement-verify loops, seven seeded workflows, live graph widget. Bring your model key.

NewNew1.0 MB snapshotStarter VM

What's inside

Harness

Pi

Plan

Starter

vCPU

1

Memory

2 GiB

Snapshot

1.0 MB

Setup prompt + walkthrough

The Pi Workflows Control Flow setup prompt + walkthrough

Paste this into any capable agent and it walks you through the connections, the trigger, and a supervised first run — the whole setup, in your own words. Or skip the setup and run this exact agent hosted.

Set up an agent for me, in its own chat, that does the work of a Pi Workflows Control Flow. Work from the files and facts I actually give it, rank what matters by impact rather than by how many items it found, and stop short of sending, filing, posting, or changing anything live unless I ask for that in the same run. Walk me through connecting Anthropic, Gemini, and OpenAI. Have it run whenever I hand it the work, and every weekday morning on whatever I left unfinished. Ask me where the source material lives, who should receive the output, and what good enough looks like for a first pass. Run one pass while I watch, then save it.

Already built, running in ~3 min · 30-day money-back on first purchase

How it works · ~3 minutes

  1. 01 · Fork

    New isolated microVM on your subdomain — creator agent state included.

  2. 02 · Your keys

    Log into Pi with your own model credentials (BYOK).

  3. 03 · Ask it to work

    Open the terminal and give the agent a real job. You keep what ships.

Example first ask

Pull my open PRs, summarize risk, and draft a review checklist.

30-day money-back on your first purchase · no free trial · keys never leave the VM

About this template

Stop leaving Pi to get control flow

You already have the long conversation going.

What you want next is a branch: propose → holy grail → y/n → implement. Not an export to n8n. Not a CI runner. Not a second harness holding half your context.

This fork boots Pi with osolmaz/pi-workflows already installed and the example graphs seeded. Type /workflow and the boxed widget is there.

The demo that proves it

export ANTHROPIC_API_KEY=...   # or OPENAI_API_KEY / GEMINI_API_KEY
pi
/workflow
/workflow echo summarize this repository

Or talk through a real problem first, then run:

/workflow elegant-solution

The graph advances above the editor: propose → holy grail → decision → implement, or a human checkpoint when it needs you. Same conversation. Same terminal.

What lands in your fork

  • The pi-workflows extensionpi install git:github.com/osolmaz/pi-workflows already run, package under ~/.pi/agent/git/…, registered in settings.json
  • Seven seeded graphs, each callable as /workflow <name>:
    • echo — the smallest end-to-end graph
    • elegant-solution — Socratic propose → holy grail → y/n → implement
    • autoimplement — implement → verify → review loop
    • branch, two-turn, shell, autoresearch
  • Runtime deps already in persist (jiti, typebox), so imports resolve on the build VM instead of dying on the first run
  • A native in-Pi widget — the graph lives where you are typing, not in a separate web viewer

What you must supply (BYOK)

One model key, exported in your own shell:

export ANTHROPIC_API_KEY=...
# or OPENAI_API_KEY / GEMINI_API_KEY
pi

Keys are never part of the template. Run /workflow list to see every seeded graph.

Your first five minutes

Fork it, export a key, run /workflow echo summarize this repository. That proves the wiring end to end in about a minute.

Then bring a real decision and run /workflow elegant-solution. Read the graph. Answer y or n. That is the whole product.

Honest limits of in-harness workflow control flow

This is Pi, and only Pi. Claude Code, Codex and OpenClaw users want a different fork.

It is control flow inside the harness — not a CI orchestrator, not a scheduler, not a job queue. Writing a graph of your own means writing TypeScript. And BYOK means what it says: you hold the model keys, and the model spend is yours.

Provenance

Built on osolmaz/pi-workflows by Onur Solmaz. Pi: pi.dev.

Fork it. Own the box. Keep the conversation. Your first purchase carries a 30-day money-back guarantee.

Inside this fork

Forking copies this template into a brand-new, fully isolated microVM on your own subdomain. Here's exactly what lands in it.

  • Pi agent

    The upstream harness, pre-installed — same version the creator ran.

  • Starter VM

    1 vCPU · 2 GiB RAM · 10 GiB disk.

  • Creator's /persist data

    The captured persist volume is copied byte-for-byte into your fork.

  • BYOK — your keys, your VM

    Add your model API keys after forking; they live only inside your microVM.

What you'll configure after forking

Secrets are scrubbed from shared templates — these are the names you supply in your agent's terminal once it boots.

Environment variables

  • ANTHROPIC_API_KEY
  • GEMINI_API_KEY
  • OPENAI_API_KEY

Your turn

Your own Pi Workflows Control Flow, live in about 3 minutes.

Copy the setup prompt and build it yourself — or let this exact Pi agent run hosted, already built, on a brand-new, fully isolated microVM on your own subdomain. Add your own keys after it boots; they never leave the box. Don't love it? Your first jurniti purchase comes with 30 days to get every cent back.

New paid VM · BYOK · 30-day money-back on your first purchase · ~3 min to provision

Needs 3 of your own API keys