Live data

SaaS Usage Limits & True Cost Directory

Look up the usage limits of automation, AI API and cloud tools before you hit the cost ceiling: rate limits, billing units and overage traps, copied from each vendor's own page on the check date.

6 tools. Last checked October 30, 2026. JSON · AI plan limits · Change log

Limit normalizer

The same workload costs a different number of units on each tool. Enter your monthly volume and see what it turns into, and which published tier covers it.

Workflow runs per monthAverage steps per workflow (trigger included)

Zapier

tasks per month

Each action step after the trigger counts as one task; Filter, Paths and Formatter are free. Cheapest tier across Pro and Team.

Make

credits per month

Every module including the trigger is one credit (routers free). Core plan, monthly billing; Pro and Teams cost more at each volume.

n8n

executions per month

One run is one execution regardless of steps. Cloud tiers as listed (annual billing, EUR); Pro has larger tiers on n8n's slider that are not listed here.

Prices are the vendor's listed figure for the smallest published tier that covers the volume; AI steps, add-ons and enterprise terms are not included. Each tool defines its unit differently, see the matrix below.

Limits matrix

Filter by category or search by name. Open a row's gotchas for the timeouts and payload caps.

Tool	Billing unit	Free allowance	Entry paid tier	Overage policy	Verified
ZapierAutomation	Per TaskA task is counted whenever Zapier successfully completes a unit of work. Triggers, Filter, Paths and Formatter steps do not count; failed actions do not count; some AI actions cost more than one task.	100 tasks/month, two-step Zaps only, 15-minute polling	Pro · $29.99/mo ($19.99/mo billed annually)750 tasks/month, multi-step Zaps, 2-minute polling; Team starts at 2,000 tasks for $103.50 ($69 annual)	Hard stopAt the limit runs are held, not lost, until the next cycle. Optional pay-per-task bills extras at 2.5x the base rate (1.25x on annual) and still pauses at 3x the subscription.	Oct 30, 2026
100 steps per Zap Zap workflows are limited to 100 steps, including all steps within paths; each action step allows up to 1,000 fields. During a free trial only the first 30 steps run. Zapier help: Zap limits Code step runtime by plan Included runtime for a Code step: Free 1 second, Pro 30 seconds, Team 30 seconds, Enterprise 2 minutes. Zapier pricing Webhook payload cap Maximum webhook payload is 10 MB for triggers and 2 MB for Catch Raw Hook, headers included. Zapier help: Webhooks by Zapier limits Polling interval by plan Free 15 minutes, Pro 2 minutes, Team and Enterprise 1 minute. Polling record counts are set by each app, not by Zapier. Zapier pricing Flood protection Polling-trigger batches of 100 or more events are held for confirmation; replayed runs process at 1 per second (1,000 runs take a little over 15 minutes). Zapier help: Zap limits Rate limits Instant triggers return 429 above 20,000 requests every 5 minutes per user. Private apps get 100 requests per 60 seconds on Free and Pro, 5,000 on Team and Enterprise. Polling triggers on Free or trial plans are held above 200 requests every 10 minutes per Zap. Zapier help: Zap limits Sources: Zapier pricing · Zapier help: Zap limits · Zapier help: Webhooks by Zapier limits
MakeAutomation	Per CreditCredits replaced operations on 27 August 2026. Every module action (read, search, create, update, delete, transform, aggregate, iterate) consumes one credit; some AI Provider features consume more. Router and error-handler modules are free.	1,000 credits/month, 15-minute minimum interval between runs	Core · $10.59/mo ($9/mo billed annually)10,000 credits/month, unlimited active scenarios, scheduling down to the minute; Pro $18.82 ($16 annual), Teams $34.12 ($29 annual) at the same volume	Hard stopScenarios stop running until credits are added; webhooks queue up to your webhook queue storage. Buy extra credits in bundles of 1,000 or 10,000, or enable auto-purchase of 10,000 credits (Core, Pro, Teams).	Oct 30, 2026
Credits expire at the end of the term Unused credits expire at the end of the month or year. Extra credits expire after one month on monthly subscriptions, or at the end of the year on annual Pro and Teams subscriptions. Make pricing Usage allowance scales with credits Per 10,000 credits licensed per month: 5 GB data transfer, 10 MB data storage, 10 MB incomplete-executions store (2 GB maximum), webhook queue of 667 (10,000 maximum). Make pricing Maximum scenario execution timenot published Not on the pricing page. Make staff on the community forum cite a 40-minute maximum with 5 minutes of grace; long jobs are split into chained scenarios. Make pricing Minimum scheduling interval 15 minutes on Free; Core and above schedule "down to the minute". Make pricing Per-credit price falls with volume Core monthly: $1.06 per 1,000 credits at 10,000; $0.94 at 20,000; $0.85 at 40,000; $0.81 at 80,000; $0.78 at 150,000; $0.71 at 300,000. Make pricing Downgrades stop scenarios Active scenarios that use features of a higher plan are stopped when you downgrade; the card is charged immediately and the billing cycle resets. Make pricing Sources: Make pricing
n8nAutomation	Per ExecutionOne workflow execution is one unit however many steps it has. Cloud plans are priced in EUR and shown with annual billing only; the self-hosted edition is free on GitHub.	No free cloud plan. Trial: 1,000 executions, 5 concurrent, 180-second timeout. Self-hosted community edition is free.	Starter · EUR 20/mo billed annually2,500 executions/month, unlimited active workflows, 5 concurrent executions; Pro EUR 50 for 10,000 executions, 20 concurrent	Auto-chargeStarter and Pro list overage fees as "Capped" without a rate. Business and Enterprise keep running and may bill overage if you do not move to the next tier.	Oct 30, 2026
Execution timeout by plan Starter 5 minutes, Pro 40 minutes, Enterprise 40 minutes; the trial allows 180 seconds. n8n pricing Concurrency Starter 5 concurrent executions, Pro 20 (the plan card says up to 50), Enterprise 200+. n8n pricing Prices are EUR, annual only The pricing page lists EUR 20, 50 and 667 per month billed annually for Starter, Pro and Business; no monthly-billing or USD figure is shown. n8n pricing Business is self-hosted only Business (40,000 executions, EUR 667/mo annual) is sold for self-hosting; cloud is waitlisted. n8n pricing Self-hosted timeout defaults EXECUTIONS_TIMEOUT defaults to -1 (off); EXECUTIONS_TIMEOUT_MAX defaults to 3600 seconds, the most a user can set per workflow. n8n docs: execution environment variables Saved-execution limits are separate Reaching the saved-execution limit does not stop workflows; it only trims how much history is kept. n8n pricing Sources: n8n pricing · n8n docs: execution environment variables
OpenAI APIAI models & APIs	Per TokenBilled per million input, cached-input and output tokens, with different rates per model, context length (over 272K input tokens costs more) and speed tier (Standard, Fast, Ultrafast); Batch is about half the Standard rate.	Free tier only in allowed geographies, capped at $100/month of usage; no free token grant listed	Build tier (prepaid credits) · $5 in credit purchases$500/month usage cap, 5,000 RPM and 1,000,000 TPM for Astra/Sol/Terra models; gpt-6.1-sol $2 in / $10 out per 1M tokens, gpt-6-luna $0.10 / $0.50	Hard stopPrepaid: requests fail when the credit balance or the tier's monthly cap is reached (429). Tiers rise automatically with cumulative purchases: Launch at $100 ($5,000/mo), Grow at $500 ($200,000/mo).	Oct 30, 2026
Failed requests still count Unsuccessful requests contribute to your per-minute limit, so tight retry loops make a 429 worse. Retry with exponential backoff and jitter. OpenAI rate limits guide slow_down on fast ramps A 429 with slow_down can occur within RPM/TPM limits when traffic rises too quickly; above 1M input TPM, grow by no more than 50% every 15 minutes. OpenAI rate limits guide Long context costs double gpt-6-astra is $10 in / $50 out per 1M tokens up to 272K input tokens and $20 / $75 above it; Fast tier is $20 / $100. OpenAI API pricing Rate limits by tier Build 5,000 RPM / 1M TPM; Launch 10,000 RPM / 4M TPM; Grow 15,000 RPM / 40M TPM (Luna models: 2M, 10M, 180M TPM). OpenAI rate limits guide Promotional pricing has an end date GPT-5.6 Sol's promotional pricing ($4 in / $20 out) is available at least through November 21, 2026. OpenAI API pricing Regional uplift Regional processing (data residency) and FedRAMP endpoints carry a 10% uplift. OpenAI API pricing Sources: OpenAI API pricing · OpenAI rate limits guide
Anthropic APIAI models & APIs	Per TokenBilled per million input and output tokens per model, with cache writes at 1.25x (5-minute) or 2x (1-hour) input and cache reads at 10% or less; Batch API is 50% off both directions.	No free tier listed; new organizations start in an Evaluation tier with lower limits until usage history builds	Start tier (prepaid credits) · Prepaid credits, $500/month spend cap1,000 RPM, 2M input and 400K output tokens/min on Sonnet, Opus and Haiku 5.5 (Fable 5.x: 500K / 100K); Sonnet 5.5 $2 in / $10 out per 1M tokens, Haiku 5.5 $0.10 / $0.50	Hard stopAt the tier's monthly spend cap (Start $500, Build $1,000, Scale $200,000) API usage pauses until 00:00 UTC on the 1st of next month unless a higher tier is granted; the 429 carries no retry-after header.	Oct 30, 2026
Spend-cap 429 looks like a rate limit The error type is rate_limit_error with error_code enforced_spend_limit_reached and no retry-after header; SDK auto-retries keep failing until access resumes. Claude Platform: rate limits Cached reads do not count toward ITPM For most models only uncached input tokens count toward input-tokens-per-minute, so prompt caching raises effective throughput (Haiku 3.5 is the exception). Claude Platform: rate limits Fable has its own, lower limits Fable 5.x shares one bucket across 5.1 and 5: 500K ITPM / 100K OTPM on Start, 1.5M / 300K on Build, 4M / 800K on Scale. Claude Platform: rate limits Haiku price doubles past 100K tokens Haiku 5.5 is $0.10 in / $0.50 out per 1M tokens for prompts up to 100,000 tokens and $0.50 / $2.50 above that. Claude Platform: pricing Acceleration limits A sharp increase in usage can return 429 even under the published limits; increase traffic gradually. Claude Platform: rate limits Batch queue caps Batch API: 1,000 RPM and 200,000 requests in the processing queue on Start, up to 100,000 requests per batch. Claude Platform: rate limits Sources: Claude Platform: pricing · Claude Platform: rate limits
SupabaseDatabases	Per ResourceA flat plan fee plus metered database disk, egress, file storage, monthly active users, Edge Function invocations and compute instances, each with its own included quota and per-unit price.	500 MB database, 5 GB egress, 1 GB storage (50 MB max upload), 50,000 MAUs, 500,000 Edge Function invocations; 2 active projects	Pro · from $25/mo8 GB disk, 250 GB egress, 100 GB storage, 100,000 MAUs, 2M Edge Function invocations, $10/month compute credit (one Micro instance)	Auto-chargeSpend cap is on by default on Pro: usage beyond quota is blocked until you turn it off, then extra usage bills at month end ($0.125/GB disk, $0.09/GB egress, $0.0213/GB storage, $0.00325/MAU, $2 per 1M invocations).	Oct 30, 2026
Free projects pause after 1 week Free projects are paused after 1 week of inactivity and limited to 2 active projects; paid plans are never paused for inactivity. Supabase pricing Second project pays full compute The $10/month compute credit covers one Micro instance; each additional project pays its full compute cost (Micro $10, Small $15, Medium $60, Large $110 per month). Supabase pricing 50 MB upload cap on Free Free file storage is 1 GB with a 50 MB maximum upload size. Supabase pricing Compute scales to $3,730/month Dedicated instances run from Large (2 vCPU, 8 GB, $110) to 16XL (64 vCPU, 256 GB, $3,730); above that contact sales. Supabase pricing Team starts at $599 Team is from $599/month with the same per-unit quotas and overage prices as Pro. Supabase pricing Sources: Supabase pricing

How this is kept

Every figure is copied from the vendor's pricing page or help centre on the check date and carries its source link. Where a vendor states no number (an execution-time cap that only appears in a community thread, for example) the row says "not published" and quotes how the vendor describes it instead of guessing. Prices in EUR stay in EUR. Reuse the JSON with a link back (CC BY 4.0).