Market pricing
Providers set their own prices and compete for requests.
The open
market
for
AI inference.
One API connects applications and agents to independent compute providers competing to serve AI inference.
OpenAI compatiblex402 nativeUSDC settlement
How it works
One API.
A market behind it.
Katara connects AI demand with available compute.
Independent providers run approved model bundles, publish their own prices, and compete for requests.
Katara routes the work, verifies the provider, and settles payment. The customer gets one API. Underneath it is a market.
Learn how the market worksApplicationsWeb apps, services, tools
AI agentsAutonomous software
Routing · Pricing · VerificationProvider 1$0.03 / 1M tokens
Provider 2$0.04 / 1M tokens
Provider 3$0.05 / 1M tokens
Verified result
USDC settlement
x402 payments
Software can pay for inference directly over HTTP.
Verified models
Versioned bundles define exactly what providers serve.
Built for the agent economy
Built for software
that spends money.
Agents already know how to call APIs. Katara gives them a way to pay for inference while they work.
Give software access to funds and spending rules. It can buy inference as it needs it.
Give your agent a budget,
not another billing account.
HTTP + x402Payment flow
- 01Request inferenceHTTP
POST /v1/chat/completions - 02Receive payment terms402
402 Payment Required - 03Authorize paymentx402
PAYMENT-SIGNATURE - 04Run inference. Return a result.
200 OK
Payment becomes part of the request.
For developers
One API.
Real choice.
Use the OpenAI client you already have. Point it at Katara and access supported models through the same request format.
Katara handles provider selection, routing, validation, metering, and settlement. You get the response.
- OpenAI compatible
- Pay per request
- Provider competition
- Managed fallback for supported models
POST /v1/chat/completionsExample
model
katara/llama-3.1-8b-instruct@1Response
usage
- prompt_tokens
- 18
- completion_tokens
- 52
- total_tokens
- 70
katara
- cost_usdc
- 0.000018
- provider
- 0xfFe5…38DC
- fallback
- false
Illustrative response from the API documentation.
For compute providers
Provide compute.
Run supported models. Set your price.
Serve requests. Earn USDC.
Katara handles demand, routing, model definitions, validation, receipts, and settlement.
- Choose approved models
- Set your own prices
- Get paid in USDC
No Katara token. No staking. No mining game.
Just compute for payment.
Provider overviewExample
Model
Llama 3.1 8B
Bundle version @1Your price / 1M tokens
- Input
- $0.03
- Output
- $0.05
Status
Available
Earnings
USDC
Example offer. Providers set their own prices.