Pricing
Native Ethiopian birr billing units for Addis AI capabilities.
Addis AI bills natively in Ethiopian birr (ETB). Any USD or EUR amount shown in account tools is an indicative conversion and can change with the exchange rate.
Current Pricing
| Capability | Billing unit | Price |
|---|---|---|
| Addis Voices 2 | Generated audio | 5 ETB / minute |
| Text Generation & Translation (Input) | Input tokens | 0.3 ETB / 1,000 tokens |
| Text Generation & Translation (Output) | Output tokens | 0.8 ETB / 1,000 tokens |
| Speech to Text | Transcribed output | 3.5 ETB / 1,000 characters |
| Realtime Audio (Input) | Input tokens | 1 ETB / 1,000 tokens |
| Realtime Audio (Output) | Output tokens | 4 ETB / 1,000 tokens |
Addis Voices 2 estimates
Use voice.estimate before generation when a user must approve spend. The estimate reports the expected duration, billable duration, cost, balance, currency, and whether generation can proceed.
Voice streaming billing
Streaming text-to-speech uses the Addis Voices 2 generated-audio rate above. Its HTTP and WebSocket speech output is billed by measured duration. The Realtime Audio input/output token rates apply to the separate Realtime API.
A text-based estimate can differ from the final charge. Read the confirmed usage after streaming completes, and reuse the original request ID and inputs for a recovery request to avoid another charge.
Billing Guidance
Track usage
Store returned usage metadata with your internal request ID when allocating cost by customer, workspace, or feature.
Retry safely
Reuse the same Addis Voices 2 client request ID only when retrying the same logical generation, so an uncertain response does not create duplicate billed clips.
The developer platform remains authoritative for account balance, credits, and any account-specific pricing.
Contract last verified: