formats

Bridge MD — formats your agents can write.

.workflow.md and .canvas.md are plain-text specs — hand the spec to any AI, or let an MCP-connected agent edit the file directly. Either way, Arg0 renders the result.

Two doors in

Have AI author the file — two ways.

Door 1

Paste the spec

workflow-format.md and canvas-format.md are portable specs written to be pasted into any AI — ChatGPT, Claude, Gemini. It replies with a valid file; save it into a workspace and Arg0 opens it.

spec .mdany AIfile on diskArg0 opens it

Door 2

MCP-connected agents

Arg0 hosts MCP servers — stdio, streamable-http, and SSE transports, OAuth flows, secrets resolved from env or the OS keychain. Give an agent a filesystem MCP server (or any file-writing tool) and it edits .workflow.md / .canvas.md in place — the app re-renders the same file.

agent + MCP toolsedit file in placelive re-render

Which door?

Paste, or plug in?

Both end at the same file — pick by how much loop you need around the write.

1Paste the spec2MCP-connected agent
SetupNone — copy the spec, paste, save the fileConfigure an MCP server once in Settings
IterationOne file per round-trip through chatLive loop — every edit re-renders in Arg0
ContextThe spec — nothing elseYour whole workspace — files, integrations, past runs
ReviewThe file lands fully formed — read, then enableWatch each edit land; checkpoints gate risky steps
Best forOne-shot authoring, sharing templates, web AIsIterative builds, refactors, in-app agents

Most people start at Door 1 for a first draft — then move to Door 2 when the file starts changing every day.

Accuracy & cost

Fewer tokens, or fewer mistakes?

The honest trade-off: Door 1 spends less per file. Door 2 spends more — and self-corrects.

Token cost — Door 1 wins

tokensnodes in graphpaste — one prompt, fixed-ishagent loop — scales with the graph

One spec in, one file out — no tool calls, no re-reads. The agent loop pays for every read, edit, and parse check.

Accuracy — Door 2 wins

correctnessedit-check passespaste — one shot, errors ship silentagent — parse-check → repair

A paste can’t verify itself — a dangling edge ships until Arg0 rejects it. The loop catches errors before runtime.

Door 1 — paste the specDoor 2 — MCP agent

The frugal pattern: draft in Door 1, then let a Door-2 agent parse-check and repair — you spend loop tokens only where errors would have shipped.

The formats

Small enough to paste, strict enough to run.

Both specs fit in one Markdown page — sized so an AI can hold the whole grammar in context and emit a valid file in one shot.

.workflow.md

— automation graph
daily_digest.workflow.md
1=== Workflow: Daily digest ===
2
3---
4name: Daily digest
5enabled: true
6---
7
8## @automation daily_digest
9
10workflow:
11 trigger: cron_9am
12 nodes:
13 - id: cron_9am
14 type: trigger.cron
15 parameters: { schedule: "0 9 * * *" }
16 - id: post
17 type: action.slack.post
18 credentials: slack
19 connections:
20 - cron_9am -> post
trigger.*http.requestcode.javascriptai.promptloop.*filterswitchmerge.*delay.*transform.*checkpoint.manualaction.<service>.<op>

.canvas.md

— infinite whiteboard
launch_board.canvas.md
1---
2title: Launch board
3type: canvas
4---
5
6## @canvas main
7
8canvas:
9 background: dots
10 nodes:
11 - id: n1
12 type: sticky
13 x: 80 y: 80
14 text: "Ship v1.1"
15 color: yellow
16 - id: n2
17 type: card
18 x: 340 y: 60
19 title: "Release flow"
20 edges:
21 - { from: n1, to: n2, style: elbow }
cardstickytextrectpillellipsediamondimagefilelinkframegroupink

Why plain text

Built to be written by anything.

No SDK required

If a model can write Markdown, it can author for Arg0. No client library, no auth dance — the file is the API.

Diffable & reviewable

An agent's edit is a git diff like any other — review it, revert it, branch it. Nothing hides in a binary blob.

One source of truth

The same file drives the canvas UI, the executor, and the docs — an agent's write and your click edit land in the same place.