API Cost Estimator
Estimate monthly API costs and cost per call from request volume, infra cost per 1k calls, vendor cost per 1k, and fixed overhead.
Inputs
Scenarios
Applies to the selected input only; adjust other inputs manually if needed.
Results
Estimated monthly total cost
-
-
Cost per call
-
-
Cost per 1,000 calls
-
-
Estimated monthly variable cost
-
-
Insights
Auto-generated from your inputs.
Adjust inputs to see recommendations.
Compare
Save a baseline to see deltas for every output.
Estimated monthly total cost
Baseline -
Delta -
Cost per call
Baseline -
Delta -
Cost per 1,000 calls
Baseline -
Delta -
Estimated monthly variable cost
Baseline -
Delta -
Sensitivity
Adjust the input to see how outputs respond to small changes.
Estimated monthly total cost
Low -
Base -
High -
Cost per call
Low -
Base -
High -
Cost per 1,000 calls
Low -
Base -
High -
Estimated monthly variable cost
Low -
Base -
High -
Guide
This page is a calculator first, but it's also a quick reference you can share internally. Start with the presets, then adjust inputs and copy the share link. Example defaults for this tool are shown below.
Example Inputs Outputs How it works Modeling tips Validation checks Common mistakes Interpretation Use cases Mini walkthroughs Scenarios Edge cases FAQ
Example (defaults)
Example inputs: API calls per month = 5000000, Infra cost per 1,000 calls = 0.02, Vendor/API cost per 1,000 calls = 0.01
Estimated monthly total cost
$650.00
Cost per call
$0.00013
Cost per 1,000 calls
$0.13
Estimated monthly variable cost
$150.00
Inputs explained
| Input | Default | Notes |
|---|---|---|
| Currency | USD | Adjust to match your product assumptions. |
| API calls per month | 5000000 | Adjust to match your product assumptions. |
| Infra cost per 1,000 calls | 0.02 | Adjust to match your product assumptions. |
| Vendor/API cost per 1,000 calls | 0.01 | Optional: LLMs, third-party APIs, enrichment, email/SMS, etc. |
| Monthly fixed overhead | 500 | Support, tooling, baseline infra, on-call - anything not per-call. |
Outputs explained
| Output | What it means |
|---|---|
| Estimated monthly total cost | A money value based on your selected currency. |
| Cost per call | A money value based on your selected currency. |
| Cost per 1,000 calls | A money value based on your selected currency. |
| Estimated monthly variable cost | A money value based on your selected currency. |
How it works
- Variable cost = (calls / 1,000) x (infra cost per 1,000 + vendor cost per 1,000).
- Total monthly cost = variable cost + fixed overhead.
- Cost per call = total monthly cost / calls.
Modeling tips
- Combine infra and vendor costs for a realistic per-call baseline.
- Use expected plan volume, not total system usage, for pricing scenarios.
- Include support and monitoring in fixed overhead when they scale by customer count.
- If vendor pricing is tiered, enter a blended average for your typical mix.
- Run a low-volume and high-volume scenario to see cost per call variance.
- If usage ramps, model average calls during the first 90 days.
Validation checks
- Monthly total cost should equal variable cost plus fixed overhead.
- Cost per 1,000 calls should be cost per call x 1,000.
- As calls increase, cost per call should trend toward variable unit cost.
- If calls are 0, cost per call should be 0; check inputs.
Common mistakes
- Leaving vendor or pass-through costs out of the model.
- Using peak call volume instead of average volume.
- Double counting bandwidth or storage costs elsewhere.
- Ignoring support and on-call costs that scale with usage.
- Assuming discounted tiers apply to all calls.
Interpretation
- Use cost per call as the floor for usage pricing decisions.
- If vendor costs dominate, consider pass-through pricing or minimums.
- Large fixed overhead suggests a base fee plus usage overages.
- Re-run after cost changes to keep pricing accurate.
Use cases
Vendor pass-through
Quantify vendor API costs to decide if they need a separate line item.
Cost audit
Benchmark unit cost before setting API pricing tiers.
Mini walkthroughs
Cost baseline
- Enter monthly calls, infra cost, vendor cost, and fixed overhead.
- Review cost per call and per 1,000 calls.
- Use the output as a pricing floor.
Vendor sensitivity
- Increase vendor cost per 1,000 calls.
- Observe cost per call change.
- Decide if pass-through pricing is needed.
Scenarios
Infra-only API
Set vendor cost to 0 to estimate pure infra cost per call.
Vendor-heavy API
Increase vendor cost per 1,000 calls to model third-party pass-through costs.
High-volume API
Scale calls to 10M+ to check if cost per call stabilizes.
Ramp-up volume
Use lower call volume to estimate early-stage unit cost.
Edge cases
- If calls per month are 0, cost per call will be undefined; use a minimum volume.
- If vendor cost is higher than infra cost, verify pass-through assumptions.
- If fixed overhead is large, cost per call can be misleading at low volume.
- Negative inputs are treated as 0; review your data if that happens.
FAQ
Should support costs be included?
If support scales with customer count, treat it as semi-variable and include a portion in fixed overhead for pricing decisions.
Why do I need vendor cost per 1,000 calls?
Many APIs have pass-through costs (LLMs, enrichment, third-party APIs). Including them prevents margin surprises.
How should I model spiky traffic?
Use a few presets (p50 vs p90 monthly calls) and compare the cost per call under each scenario.
Do I include bandwidth or storage here?
Only if they scale with calls and you want a single blended unit cost; otherwise model them separately.
Can I use this for non-API usage?
Yes. Replace calls with any repeatable unit (jobs, events, messages) as long as costs scale with that unit.