AI workflow automation cost in USD: worked examples
AI Flowi Workflow has five plans: Go, Plus, Pro 100, Pro 200 and Pro 500. For connected workflows, billed monthly, a six-node request uses 600 tokens at 100 requests (Go, USD10), 6,000 at 1,000 (Plus, USD20) and 60,000 at 10,000 (Plus with five top-ups, USD170 = 20 + 5 × 30). A top-up adds another month’s allowance (Go USD15, Plus USD30). Third-party API and model charges are separate. Pricing checked 1 November 2026.
By Ai Flowi · Published .
What you pay for
AI Flowi Workflow is a visual workflow builder from Ai Flowi (aiflowi.com, operated by AI FLOWI SDN. BHD.) for building AI agents, chatbots and automations from drag-and-drop components. Go is USD8/month billed yearly (USD96/year) or USD10/month billed monthly. Plus is USD16/month billed yearly (USD192/year) or USD20/month billed monthly. Pro 100 is USD80/month billed yearly (USD960/year) or USD100/month billed monthly. Pro 200 is USD160/month billed yearly (USD1,920/year) or USD200/month billed monthly. Pro 500 is USD400/month billed yearly (USD4,800/year) or USD500/month billed monthly.
Each node run uses 1 Flowi Token. Tokens included each month: Go: 5,000; Plus: 10,000; Pro 100: 50,000; Pro 200: 100,000; Pro 500: 250,000. Top-ups: when a month’s tokens are used up, pay the top-up price for another month’s allowance: Go USD15 for 5,000 tokens; Plus USD30 for 10,000 tokens; Pro 100 USD150 for 50,000 tokens; Pro 200 USD300 for 100,000 tokens; Pro 500 USD750 for 250,000 tokens.
- Workflow subscription: Go costs USD10/month billed monthly and includes 5,000 tokens; Plus costs USD20 and includes 10,000. Read the pricing page.
- Third-party API and model usage: not included. Budget these separately using your provider’s charges and actual usage. A Flowi Token is a node-run unit; it is not a language model’s input or output token.
- Usage above your plan’s monthly tokens: buy a top-up: another whole month’s allowance (Go USD15, Plus USD30, Pro 100 USD150, Pro 200 USD300, Pro 500 USD750). See pricing or contact Ai Flowi for options beyond the included tokens.
For an estimate within a plan’s allowance, the monthly-billed subscription costs that plan’s price (Go USD10, Plus USD20), plus third-party charges. A no-model workflow avoids a model call; it does not establish that every external service is free.
How to count node runs
Count each time a node executes, including the trigger, validation, routing, processing, action and response nodes on the executed path. A node placed on the canvas but skipped by that path is not an additional execution. The workflow basics guide explains components and connections; count the configured path rather than its connections.
Monthly Flowi Tokens = total node runs across all workflow executions that month. If every request follows the same path once, with no retries or loops: tokens = nodes executed per request × requests per month.
Branching: count the selected path
The If-Else tutorial documents True and False outputs. The tested Sheets recipe uses Verification Router with separate valid and rejected responses. Count the router and the nodes that execute on the chosen branch. For a mixed workload, calculate each path separately and add the results.
For example, assuming 800 requests follow the six-node valid path and 200 follow the four-node rejected path: (800 × 6) + (200 × 4) = 5,600 Flowi Tokens. This fits Plus’s 10,000-token allowance and costs USD20/month billed monthly before third-party charges; see pricing or contact Ai Flowi. This is calculated usage, not a measured monthly bill.
Retries: count repeated node executions
A retry that executes nodes again adds those node runs. If a sender repeats the entire six-node valid path 50 times, those extra deliveries add 50 × 6 = 300 tokens. If only part of a flow re-executes, count that part. This is a budgeting assumption, not a claim that automatic retries are available.
The 8 October append recipe writes another row for an identical delivery. The version tested on 9 October prevents a second row for sequential repeats by updating the matching row; its valid path has seven nodes, including response projection. It still does work for the repeated request. Simultaneous duplicates can both insert. In version 1.9.2, a Sheets failure stops the flow and returns the raw platform error; no in-flow error branch catches it.
Loops: multiply the repeated work
The component catalog documents Loop; the use-case guide describes a Research Translation Loop that translates search results. A loop can execute the same nodes for multiple items, so count executions per item as well as all setup, loop-control and finishing executions.
For illustration, three processing nodes executing once for each of 20 items contribute 3 × 20 = 60 node runs, before adding the other nodes that execute. The docs cited here do not establish a fixed number of loop-controller executions or measured loop billing. Check the actual records for your configuration.
Use Workflow usage and token history to inspect Step, Tokens and Status for recorded charges. Its Workflow Runs view counts charged steps, not distinct complete requests; Overview reads the newest 1,000 records in the selected window and may have incomplete totals for a busier window. Include other workflows and recorded charges when checking the shared allowance.
Worked scenarios: 100, 1,000 and 10,000 runs per month
Here, one run means one incoming request that completes the stated path. Every node on that path executes once. All examples assume no retries, no loops and no additional branch executions. The table estimates node usage, not throughput, production capacity or an all-inclusive bill.
Production-tested Sheets recipe: six nodes on the valid path
The 8 October 2026 recipe has seven nodes on the canvas. A valid authenticated request executes these six:
- Universal Webhook receives the request.
- JSON Operations validates the input.
- Verification Router selects Verified.
- A second JSON Operations prepares the row.
- Append Rows writes to Google Sheets.
- Webhook Response returns 200.
The separate 422 response is not on this valid path. The original and imported-copy tests recorded a 200 response and one added row for valid input; invalid input returned 422 and missing or invalid keys returned 403 without new rows. Those tests verify the recipe’s behaviour; the monthly token totals below are arithmetic estimates. This version has no model call and one Append Rows operation per valid request. Any external-service charges remain separate.
Webhook → validate → respond: three nodes
This illustrative flow has one webhook, one JSON validation/processing node and one response node. It checks an incoming request and returns a result, with no Sheets write, router, model call or other external API call assumed. All three nodes execute once per request. It is a budgeting example, not a production-tested recipe.
Larger flow: ten executed nodes
This illustrative flow assumes ten nodes execute once per request on its selected path. It includes one third-party API or model call per request: 100, 1,000 or 10,000 calls at the respective monthly volumes. The provider’s price and usage are separate inputs, so no USD total for that service is assumed. This is a sizing example, not a tested ten-node implementation.
| Workflow | Nodes per run | Runs/month | Flowi Tokens/month | Monthly cost on the cheapest plan (USD, billed monthly) |
|---|---|---|---|---|
| Tested append recipe: valid path | 6 | 100 | 600 | Go USD10 up to 5,000 tokens, Plus USD20 up to 10,000 |
| Tested append recipe: valid path | 6 | 1,000 | 6,000 | Go USD10 up to 5,000 tokens, Plus USD20 up to 10,000 |
| Tested append recipe: valid path | 6 | 10,000 | 60,000 | USD170 on Plus = 20 + 5 × 30 |
| Webhook → validate → respond | 3 | 100 | 300 | Go USD10 up to 5,000 tokens, Plus USD20 up to 10,000 |
| Webhook → validate → respond | 3 | 1,000 | 3,000 | Go USD10 up to 5,000 tokens, Plus USD20 up to 10,000 |
| Webhook → validate → respond | 3 | 10,000 | 30,000 | USD80 on Plus = 20 + 2 × 30 |
| Illustrative larger flow | 10 | 100 | 1,000 | Go USD10 up to 5,000 tokens, Plus USD20 up to 10,000 |
| Illustrative larger flow | 10 | 1,000 | 10,000 | Go USD10 up to 5,000 tokens, Plus USD20 up to 10,000 |
| Illustrative larger flow | 10 | 10,000 | 100,000 | USD200 on Pro 200 (100,000 tokens included) |
Monthly cost = the plan’s monthly price + number of top-ups x its top-up price. A top-up adds another whole month’s allowance of that plan, so round the number of top-ups up. Billed monthly: Go USD10 (5,000 tokens, top-up USD15), Plus USD20 (10,000 tokens, top-up USD30), Pro 100 USD100 (50,000, USD150), Pro 200 USD200 (100,000, USD300), Pro 500 USD500 (250,000, USD750). Check the pricing page or contact Ai Flowi for options beyond the included tokens. Each row assumes this is the only workload consuming your plan’s allowance. Third-party charges are separate.
Competitor context
- AI Flowi Workflow and n8n comparison
- AI Flowi Workflow and Dify comparison
- Make and AI Flowi Workflow comparison
- Zapier and AI Flowi Workflow comparison
Estimate your workflow
Use the Flowi Token usage calculator with the number of nodes that actually execute, your monthly volume, longer branches and retries. Enter third-party API or model costs separately. Then compare the estimate with recorded usage before budgeting for a larger workload.