News/November 20, 2026
ClawStage credit update
Thank you for joining the ClawStage beta, and for telling us how it feels to use.
Some of you said a voice chat can run out of credits before the conversation is finished. Creating a character can cost a lot too, especially when you want to revise it a few times and try different looks. Those costs do get in the way, so this update lowers the credits used for chat and for several character edits.
The new rates, and how they are counted, are below, so you can plan your credits.
Credit rates
Voice conversation
| Step | Unit | Credits |
|---|---|---|
| LLM input | per 10,000 tokens | 12 Cr |
| LLM output | per 1,000 tokens | 6 Cr |
| ASR | per minute | 10 Cr |
| TTS | per 50 characters | 4 Cr |
Workshop character generation
| Step | Unit | Credits |
|---|---|---|
| Text generation | each time | 10 Cr |
| Portrait image | per image | 60 Cr |
| Portrait image edit | per image | 30 Cr |
| Standby image | per image | 60 Cr |
| Standby image edit | per image | 60 Cr |
| Video generation | per clip | 350 Cr |
What changed
LLM input (per 10,000 tokens)
Before 23 Cr→After 12 Cr
−48%
LLM output (per 1,000 tokens)
Before 14 Cr→After 6 Cr
−57%
ClawStage speech (per 50 characters)
Before 11 Cr→After 4 Cr
−64%
Text generation (each time)
Before 50 Cr→After 10 Cr
−80%
Portrait image (per image)
Before 54 Cr→After 60 Cr
+11%
Portrait image edit (per image)
Before 54 Cr→After 30 Cr
−44%
Standby image (per image)
Before 200 Cr→After 60 Cr
−70%
Standby image edit (per image)
Before 200 Cr→After 60 Cr
−70%
| Step | Before | After | Change |
|---|---|---|---|
| LLM input (per 10,000 tokens) | 23 Cr | 12 Cr | −48% |
| LLM output (per 1,000 tokens) | 14 Cr | 6 Cr | −57% |
| ClawStage speech (per 50 characters) | 11 Cr | 4 Cr | −64% |
| Text generation (each time) | 50 Cr | 10 Cr | −80% |
| Portrait image (per image) | 54 Cr | 60 Cr | +11% |
| Portrait image edit (per image) | 54 Cr | 30 Cr | −44% |
| Standby image (per image) | 200 Cr | 60 Cr | −70% |
| Standby image edit (per image) | 200 Cr | 60 Cr | −70% |
What this means for you
We lowered the cost of many character creation and edit steps, so you can try more looks with less worry about the balance.
Video generation still uses the Seedance model. As Seedance 1.0 is retired, we will move to Seedance 2.0 and add super-resolution, so the video is clearer and the character's details are easier to see. The new model costs us more, but the price stays at 350 credits per clip. We cover the difference.
We are also preparing a character community. Once it opens, you will be able to use characters other creators have already made, at a lower credit cost, instead of generating and revising one from scratch. We will share how it works, and what it costs, when the community opens.
Why these prices can come down
To leave more room for chat and creation, we made three changes:
Better rates from our providers. We renegotiated our agreements and committed to a larger volume of service. That lowered the cost of each use, and the saving is reflected in the credit prices.
Models that look better and cost less. As new models arrived, we compared quality and price again and replaced some of them, so the same task can be done for less.
A simpler workflow. We changed how chat and character generation are processed, and removed repeated steps, so the AI finishes the work more efficiently. That cuts unnecessary use, and it lowers the cost of each chat and each creation.
These changes let us lower the credit price of several features. We hope the same balance now lasts for a longer conversation, and for a few more looks you want to try.
Will prices change again?
Yes. The AI services we use will keep changing, and provider prices can go up or down. We will review these costs regularly and adjust credit prices to match what they actually cost us.
We will tell you before each change takes effect, so you have time to read it and plan how you use your credits.