Cost Estimation
The primary cost driver for RAISE is AWS Bedrock — specifically the token usage for Claude Sonnet 4.6 inference. All other services (Lambda, SQS, S3, DynamoDB, SNS, CloudWatch) operate within or very close to the AWS Free Tier at typical resume processing volumes.
You are responsible for the cost of AWS services used when running RAISE. The estimates below are based on standard AWS pricing in the us-east-1 region as of 2026 and may vary depending on your region, usage patterns, and any applicable free-tier eligibility. Always refer to the AWS Pricing Calculator for up-to-date figures.
Cost Assumptions
| Assumption | Value |
|---|---|
| Average resume size | 500 KB |
| Lambda duration per invocation | ~50 seconds (512 MB memory) |
| Bedrock input tokens per resume | ~2,000 tokens |
| Bedrock output tokens per resume | ~200 tokens |
| Claude Sonnet 4.6 input price | $3.00 / 1M tokens |
| Claude Sonnet 4.6 output price | $15.00 / 1M tokens |
| Output target | BOTH (DynamoDB + SQS) |
| S3 storage | ~0.5 GB (within free tier) |
| SQS requests | Within free tier (first 1M/month free) |
Example 1: 10 Resumes per Month
Suitable for small teams or pilot deployments.
| Service | Dimensions | Monthly Cost (USD) |
|---|---|---|
| AWS Lambda | 10 invocations × 50s × 512 MB | $0.00 |
| AWS Bedrock | 10 × 2,000 input tokens + 10 × 200 output tokens | $0.09 |
| Amazon S3 | ~0.5 GB storage, 20 requests (10 PUT + 10 GET) | $0.00 |
| Amazon SQS | 30 requests (ingestion + output + DLQ) | $0.00 |
| Amazon DynamoDB | 10 writes (on-demand billing) | $0.00 |
| Amazon SNS | Within free tier (first 1M notifications/month) | $0.00 |
| Amazon CloudWatch | Basic metrics and dashboard | $0.00 |
| Total | $0.09 |
Example 2: 100 Resumes per Month
Suitable for active hiring periods at small-to-medium organisations.
| Service | Dimensions | Monthly Cost (USD) |
|---|---|---|
| AWS Lambda | 100 invocations × 50s × 512 MB | $0.00 |
| AWS Bedrock | 100 × 2,000 input tokens + 100 × 200 output tokens | $0.90 |
| Amazon S3 | ~0.5 GB storage, 200 requests (100 PUT + 100 GET) | $0.00 |
| Amazon SQS | 300 requests (ingestion + output + DLQ) | $0.00 |
| Amazon DynamoDB | 100 writes (on-demand billing) | $0.00 |
| Amazon SNS | Within free tier | $0.00 |
| Amazon CloudWatch | Basic metrics and dashboard | $0.00 |
| Total | $0.90 |
Example 3: 1,000 Resumes per Month
Suitable for high-volume recruitment operations or enterprise talent pipelines.
| Service | Dimensions | Monthly Cost (USD) |
|---|---|---|
| AWS Lambda | 1,000 invocations × 50s × 512 MB | $0.00 |
| AWS Bedrock | 1,000 × 2,000 input tokens + 1,000 × 200 output tokens | $9.00 |
| Amazon S3 | ~0.5 GB storage, 2,000 requests (1,000 PUT + 1,000 GET) | $0.00 |
| Amazon SQS | 3,000 requests (ingestion + output + DLQ) | $0.00 |
| Amazon DynamoDB | 1,000 writes (on-demand billing) | $0.04 |
| Amazon SNS | Within free tier | $0.00 |
| Amazon CloudWatch | Basic metrics and dashboard | $0.00 |
| Total | $9.04 |
Cost Optimisation Tips
- Prompt efficiency — The Bedrock prompt is the only meaningful cost lever. Reducing the average token count per resume (e.g. by trimming verbose prompt instructions) directly reduces cost.
- Output target — If you only need DynamoDB and have no downstream SQS consumer, set
outputTarget=DYNAMODBto avoid the SQS publish cost (though SQS remains free at these volumes). - S3 lifecycle policy — Add an S3 lifecycle rule to expire raw resumes after 90 days to keep storage costs near zero as volume grows.
- Reserved capacity — At very high volumes (10,000+ resumes/month), consider AWS Budgets and Bedrock cost monitoring to track spend proactively.