Skip to main content

Detailed API Costs

This page provides detailed pricing for all API services used by SmarterAvatar.

Avatar Streaming - HeyGen LiveAvatar

HeyGen is the primary cost driver for most deployments.

Credit Pricing

TierMonthly CostCreditsPer Credit
Pro$99100$0.99
Scale$330660$0.50
EnterpriseCustomCustomNegotiated

Streaming Rates

API streaming: 1 credit = 5 minutes

DurationPro TierScale Tier
30 sec (minimum)$0.099$0.05
1 minute$0.198$0.10
2 minutes$0.396$0.20
3 minutes$0.594$0.30
5 minutes$0.99$0.50
10 minutes$1.98$1.00
Billing Details
  • Minimum charge: 30 seconds per session
  • After minimum: billed per second
  • Credits expire 30 days after issue

Speech-to-Text

OpenAI Whisper

MetricCost
Per minute$0.006
Per hour$0.36
Billing incrementPer second
Minimum chargeNone

Deepgram Nova-2

PlanPer MinutePer Hour
Pay-as-you-go$0.0077$0.46
Growth$0.0065$0.39

Free tier: $200 credit (never expires)

AWS Transcribe

If using hybrid architecture with AWS:

Volume/MonthPer Minute
First 250K min$0.024
250K - 1M min$0.015
1M - 5M min$0.0102
5M+ min$0.0078

LLM Providers

Google Gemini

ModelInput/1M tokensOutput/1M tokens
gemini-3.0-flash$0.50$3.00
gemini-2.5-flash$0.15$0.60
gemini-2.5-pro$1.25$5.00

RAG (File Search):

  • Storage: FREE
  • Indexing: $0.15 per 1M tokens (one-time)
  • Query embeddings: FREE

OpenAI

ModelInput/1M tokensOutput/1M tokens
gpt-4o-mini$0.15$0.60
gpt-4o$2.50$10.00
gpt-4-turbo$10.00$30.00

Anthropic

ModelInput/1M tokensOutput/1M tokens
claude-3-5-haiku$0.80$4.00
claude-sonnet-4$3.00$15.00
claude-opus-4$15.00$75.00

AWS Bedrock (Claude)

Same as Anthropic direct, accessed through AWS:

ModelInput/1M tokensOutput/1M tokens
Claude 3.5 Sonnet$3.00$15.00
Claude 3 Haiku$0.25$1.25

Session Cost Examples

Example 1: Budget Configuration

3-minute session, 4 interactions

ComponentCalculationCost
HeyGen (Scale)3 min × $0.10$0.30
Whisper1 min speech × $0.006$0.006
Gemini Flash~4K tokens$0.01
Total$0.316

Example 2: Premium Configuration

3-minute session, 4 interactions

ComponentCalculationCost
HeyGen (Scale)3 min × $0.10$0.30
Deepgram1 min speech × $0.0077$0.008
Claude Sonnet~4K tokens$0.08
Total$0.388

Example 3: AWS Hybrid Configuration

3-minute session, using your Bedrock

ComponentCalculationCost
HeyGen (Scale)3 min × $0.10$0.30
AWS Transcribe1 min × $0.024$0.024
Bedrock (Claude)~4K tokens$0.08
Total$0.404

Cost Optimization Tips

1. Use Scale Tier for HeyGen

Pro: $0.198/min → Scale: $0.10/min = 49% savings

2. Choose Efficient LLM

Gemini Flash: $0.01/session vs Claude: $0.08/session = 87% savings

3. Optimize Session Length

Set reasonable timeouts to avoid idle sessions burning credits.

4. Cache Common Responses

Use response overrides for FAQs to skip LLM calls entirely.

5. Monitor Usage

Use the admin dashboard to track costs and identify optimization opportunities.