CFFBRW is a workflow automation platform where you write tasks in plain English markdown. An AI compiler turns your prose into a deterministic execution plan, and CFFBRW runs it on Cloudflare's global edge — with self-healing when steps fail.

Does CFFBRW require coding?

No coding is required. You write workflow steps in plain English markdown. The AI compiler infers step types, dependencies, and execution order from your descriptions automatically.

How does CFFBRW compare to Zapier?

Zapier is a mature, respected platform for linear trigger-action automation. CFFBRW is designed for more complex workflows: AI-compiled execution graphs from prose, parallel step execution, self-healing when APIs fail, and human approval gates that can pause for days.

What AI models does CFFBRW support?

CFFBRW supports Claude, Gemini, and Qwen for workflow compilation. You can also bring your own AI agent via the BYOA (Bring Your Own Agent) feature — connect CFFBRW as an MCP server in your IDE or agent, and your AI compiles the workflows.

What happens when a workflow step fails?

CFFBRW has Tier 2 AI recovery: when a step fails, AI diagnoses the error, sees the prior step outputs, and either corrects the output directly, retries with fixed parameters, or escalates to a human. You see 'Adapted' instead of 'Failed.'

Use case

Turn raw data into weekly reports — automatically.

Pull metrics from multiple APIs every Monday. Run parallel AI analysis on each data stream. Compile into a formatted summary. Deliver to your team via Slack or email. Zero manual work.

Start building — free Read the docs

The problem

Someone spends Monday morning pulling numbers together.

Export from the CRM. Export from the analytics dashboard. Pull from the payments API. Paste into a spreadsheet. Write a summary. Format it. Copy into Slack. Every week.

This is exactly the kind of multi-step, multi-source, recurring task that should never touch a human hand.

The workflow

Write it once. Runs every Monday at 9am.

weekly-report.mdcron: 0 9 * * MON

# Weekly Business Report

### 1. Fetch revenue data

Get last 7 days of revenue from the payments API. No dependencies.

### 2. Fetch user signups

Get new user count from the analytics API. Runs in parallel with step 1.

### 3. Fetch support tickets

Get open ticket count and top categories. Runs in parallel with steps 1 and 2.

### 4. AI trend analysis

Using all three data sources, identify the top 3 business trends this week. Depends on steps 1, 2, AND 3.

### 5. Compile report

Format the data and analysis into a structured weekly report with key metrics and recommendations.

### 6. Send to team

Post the report to #weekly-updates on Slack and email to the leadership list.

Why CFFBRW

Parallel execution. AI analysis. Self-healing.

⚡

Steps 1, 2, and 3 run simultaneously

AI infers from prose that the three API fetches are independent. They run in parallel — the report takes seconds, not minutes.

AI writes the analysis, not you

The ai_call step gets all three data streams as context and produces a written summary. No templates, no string formatting.

🔁

Self-healing if an API goes down

If the payments API returns an error, AI recovery diagnoses and retries with adjusted parameters before escalating to you.

Step types used

Built with 4 step types and a cron trigger.

cron trigger — fires every Monday at 9am UTC

http_request ×3 — parallel fetches from each data source

ai_call — trend analysis with all data as context (fan-in)

transform — compiles data into report format

connector — delivers to Slack and email

Get started

Your first automated report
running this Monday.

Write the workflow in plain English. CFFBRW compiles and schedules it. No credit card. No config files.

Start building — free See more examples